Вы находитесь на странице: 1из 76

SUSTAINING PRIVACY AND OPEN JUSTICE

IN THE TRANSITION TO ONLINE COURT RECORDS:


A MULTIDISCIPLINARY INQUIRY +

AMANDA CONLEY, ∗ ANUPAM DATTA, ∗∗


HELEN NISSENBAUM ∗∗∗ & DIVYA SHARMA ∗∗∗∗

I. INTRODUCTION ................................................................................. 773


II. COURT RECORDS AND THE PRESUMPTION OF ACCESS ..................... 778
A. Definitions .............................................................................. 778
B. Legal Basis for the Right of Access............................................. 784
1. Sources of the Right of Access ............................................. 784
2. Contours of the Right of Access ........................................... 787
a. Who May Access: The Actors ........................................ 788
b. How Records Can Be Accessed at the Courthouse:
Transmission Principles.............................................. 789
c. What May Be Accessed: Information Type...................... 790
C. Balancing the Presumption of Access Against Other
Considerations ....................................................................... 797
III. CONTEXTUAL INTEGRITY ................................................................ 803

Copyright © 2012 by Amanda Conley, Anupam Datta, Helen Nissenbaum, and Divya
Sharma.
+
Acknowledgments: We owe the greatest debt to Grayson Barber, Steve Schultze, and
Tim Lee, who generously shared their considerable expertise over the long haul of this
project. Jim Rebo and Frank Hoeber of the New Jersey Administrative Office of the Courts
graciously opened their doors and provided invaluable guidance, not least by detailing for
us the incredible intricacies of the inner workings of a courthouse. We are grateful to Bob
Deyling and Thomas Clarke, whose feedback at various stages of the project much im-
proved our arguments. We are also grateful for opportunities to present this work to the
NYU Privacy Research Group, Privacy Law Scholars Conference 2011, and the Conference
on Privacy and Public Access to Court Records (2008 and 2011). Support for this project
came from the National Science Foundation, Cyber-Trust, Award No. CNS-0831124.

Associate, O’Melveny & Myers, LLP, San Francisco. M.A. 2008, University of Colo-
rado (sociology); J.D. 2011, New York University School of Law. The opinions in this Ar-
ticle are solely those of the authors and do not reflect the opinions of O’Melveny & Myers,
LLP, or its clients.
∗∗
Assistant Research Professor, CyLab, Electrical & Computer Engineering, and (by
courtesy) Computer Science Department, Carnegie Mellon University.
∗∗∗
Professor, Media, Culture, and Communication, and Computer Science, New York
University.
∗∗∗∗
Ph.D. candidate, Electrical & Computer Engineering, Carnegie Mellon University.

772

Electronic copy available at: http://ssrn.com/abstract=2112573


2012] ONLINE COURT RECORDS 773

IV. COMPARING PATTERNS OF INFORMATION FLOW ............................ 808


A. Information Retrieval Model .................................................... 810
B. Overview of Empirical Study..................................................... 814
1. Search Using Online Systems ............................................. 814
a. Systems Used................................................................ 814
b. Observations from Search Experience ............................. 815
2. Search Using Physical Systems............................................ 818
a. Systems Used................................................................ 818
b. Observations from Search Experience ............................. 820
C. Root Causes of Cost Differences in Online and Local Access to
Court Records......................................................................... 821
V. EVALUATING ACCESS PRACTICES AND POLICIES ............................... 824
A. Impacts of Information Flows in General Moral and Political
Terms .................................................................................... 824
B. Values and Purposes Internal to Courts and the Justice System ... 832
1. Ends and Purposes ............................................................ 832
2. Values .............................................................................. 833
3. Function of Court Records ................................................. 835
4. Disruptive Information Flows............................................. 836
VI. RECOMMENDATIONS FOR ACCESS ................................................... 839
VII. CONCLUSION: A HIDDEN VARIABLE .............................................. 845

I. INTRODUCTION
Courts, like other institutions, are undergoing a transformation:
from largely paper-based systems of processing and record keeping to
digital records, and from primarily locally accessible records to
records accessible online via the Internet. The adoption of new me-
dia, however, is not a new experience for courts. They have adapted
to video and audio recording, to microfilm and computer tape, and,
in the more distant past, to novel indexing schemes like citation
tables and legal citation indexes. 1 Members of the public wishing to
consult court records increasingly find a point of entry in computer
terminals located in courthouses rather than in the traditional manila

1. See generally Patti Ogden, “Mastering the Lawless Science of Our Law”: A Story of Legal
Citation Indexes, 85 LAW LIBR. J. 1 (1993) (discussing the chronology and reasoning behind
the evolution of the legal citation index, from the first comprehensive index of overruled
cases published in 1821 to online citation systems, first developed in the 1960s).

Electronic copy available at: http://ssrn.com/abstract=2112573


774 MARYLAND LAW REVIEW [Vol. 71:772

folders handed over the counter by a court clerk. 2 Yet while the pros-
pect of a move toward fully digitized records accessible online with no
gatekeepers has found no shortage of eager supporters and enthu-
siasts, it has also stirred a chorus of concern among academics and
court administrators. 3 Why is this? Of what interest, beyond its purely
technological implications, can a change in medium be? Why should
anyone outside of a court bureaucracy worry about what goes into
records and whether they can be accessed via a doorway to the court-
house or a portal on the Web? These are the questions we address in
this Article: not only why we should pay attention to these questions
and concerns, but how to go about answering them and crafting
access policies accordingly.
Court records exist at the confluence of two strong currents in
liberal democratic societies. One current is the demand for open-
ness. Because records provide an essential window into the function-
ing of one of the three pillars of government—the courts—citizens
are presumed to have a right to inspect them to ensure that courts are
exercising their powers not only competently and fairly but also with-
in the limits of their mandate. 4 The other current is privacy. 5 The
courts are a stage where many of life’s dramas are performed, where
people may be shamed, vindicated, compensated, punished, judged,
or exposed. These human dramas are chronicled through court
records, which include volumes of information about the various
people involved in a given dispute. It is only natural, and to be ex-

2. See Nancy S. Marder, From “Practical Obscurity” to Web Disclosure: A New Understanding
of Public Information, 59 SYRACUSE L. REV. 441, 444 (2009) (explaining that, before the In-
ternet, an individual wishing to consult court records “would have to go down to the
courthouse, request the document, examine it at the courthouse or perhaps copy each
page, and then examine the copy at home”).
3. See Daniel J. Capra et al., The Philip D. Reed Lecture Series, Conference on Privacy and
Internet Access to Court Files, Panel One: General Discussion on Privacy and Public Access to Court
Files, 79 FORDHAM L. REV. 1, 1–22 (2010) (providing the views of certain academics and
court professionals attending a conference on public access to court files and the asso-
ciated privacy concerns).
4. See Peter W. Martin, Online Access to Court Records—From Documents to Data, Particu-
lars to Patterns, 53 VILL. L. REV. 855, 857–58 (2008) (noting several reasons that justify pub-
lic access to court records, including enforcing public confidence in the judicial system,
assuring the fairness of judicial proceedings, and permitting a public check on the courts).
5. See Reena Raggi, The Philip D. Reed Lecture Series, Conference on Privacy and Internet
Access to Court Files, Welcome and Opening Remarks, 79 FORDHAM L. REV. 1, 2 (2010) (speak-
ing about the traditional concerns that have been raised about disclosure of personal in-
formation in court files).
2012] ONLINE COURT RECORDS 775

pected, that the creation and exposure of these accumulated volumes


raise privacy concerns. 6
Striving to reconcile these crosscurrents, a complex body of
rules, regulations, principles, and policies govern the creation of
court records and access to them. 7 In the United States, a range of
sources, from constitutional principles 8 to the material facts of partic-
ular cases, 9 shape these constantly evolving sets of constraints at the
federal, state, and local levels. 10 Some commentators warn that rules
and procedures developed for locally accessed or hardcopy records
cannot directly be transferred over to electronic records. They also
argue that changes brought by new media disrupt a delicate balance
by tipping the scale in favor of openness and unacceptably against
privacy. 11 Because of these substantive ramifications, the debate over
a medium becomes one about core societal values, and it spills
beyond court bureaucracies. Hence we find that academics, court
administrators, judges, public interest advocates, and citizens are
among those calling for discretion and study. 12

6. See Capra, supra note 3, at 4–5 (Joel Reidenberg saying that the Internet removed
much of the obscurity surrounding court records and, as such, that this open access has
raised privacy and public safety implications).
7. See, e.g., Copeland v. Copeland, 966 So.2d 1040, 1050–51 (La. 2007) (noting the
state legislature’s power to make exceptions to the right of public access to court records,
the individual’s power to challenge the disclosure of specific court records, and the court’s
power to determine whether a privacy interest exists in what those records contain, and
noting that most states can close or limit access to court records in specific instances).
8. See infra notes 58–63 and accompanying text (discussing the lack of clarity sur-
rounding a constitutional right to access court records).
9. See, e.g., infra text accompanying note 84 (listing certain kinds of cases where, in
New Jersey, records are excluded from public access).
10. See infra Part II.B.2.
11. See Capra, supra note 3, at 4–6 (Joel Reidenberg arguing that the procedure for ac-
cessing physical records effectively protected privacy but that electronic records implicate
privacy concerns, and suggesting changes that could protect privacy in an online court
record system).
12. A yearly conference held in Williamsburg, Virginia, for example, offers scholars,
practitioners, judges, and policy advocates the opportunity to discuss the move from paper
to online court records. See Eighth Conference on Privacy & Public Access to Court
Records, held by Center for Legal & Court Technology (Nov. 3-4, 2011),
http://www.legaltechcenter.net/education/conferences/8th-conference-on-privacy-
public-access-to-court-records/ (providing the proposed itinerary and topics for discussion
for the November 2011 conference). It bears noting that not all participants at the confe-
rence are necessarily against full and open access to court records on the Internet, and not
all favor privacy over judicial openness. Members of the Reporters Committee for the
Freedom of the Press, for example, strongly advocate for entirely open and accessible
court records. See Capra, supra note 3, at 12–16 (Lucy Dalglish discussing how open court
records are beneficial to journalists).
776 MARYLAND LAW REVIEW [Vol. 71:772

Taking on questions about the disruptive potential of electronic


media and digital networks, this Article aims to contribute to the de-
bate over what courts ought to do. At this time, avoiding electronic
media and digital networks entirely is not a serious option, nor, in our
view, is it defensible. Some readers, however, might wonder whether
anything less than full adoption is even worth considering, given that
the federal courts have already committed to the Public Access to
Court Electronic Records (“PACER”) 13 online record system. In our
view, there remain issues of principle both in PACER and other elec-
tronic access systems that have not been adequately addressed. 14 Our
Article and its concluding recommendations focus primarily on state
and local, rather than federal, courts. We do this in part because pub-
lic and internal deliberations over state access policies have remained
actively in progress in the period during which we conducted research
and wrote this Article. 15 But, just as importantly, we focus on state
court records, particularly at the trial level, because they contain an
abundance of personal information, some of which may drop away as
cases move from trial courts to appellate courts. 16 Because of the

13. PACER is an electronic public access service provided by the Administrative Office
of the U.S. Courts that contains “case and docket information from federal appellate, dis-
trict and bankruptcy courts, and the PACER Case locator via the Internet.” PUBLIC ACCESS
TO COURT ELECTRONIC RECORDS, http://www.pacer.gov (last visited Mar. 8, 2012).
14. For example, Peter Martin makes the important and often overlooked point that
while one purported purpose of increased access to court records is to enhance judicial
accountability, PACER—a system developed for use by lawyers and judges, rather than by
the general public—does not allow users to search cases by judge. See Martin, supra note 4,
at 871 (“[T]here is no search feature comparable to PACER’s party name search that
would allow a user to gather and inspect judge or attorney actions across multiple cases.
Of course, the system holds this data, but it does not permit the data fields for judges and
attorneys to be the subjects of search.” (footnote omitted)). This is not PACER’s only limi-
tation. See, e.g., John Schwartz, An Effort to Upgrade a Court Archive System to Free and Easy,
N.Y. TIMES, Feb. 12, 2009, at A16 (“‘Pacer is just so awful,’ said Carl Malamud, the leader
of the effort and founder of a nonprofit group, Public.Resource.org. ‘The system is 15 to
20 years out of date.’ Worse, Mr. Malamud said, PACER takes information that he believes
should be free—government-produced documents are not covered by copyright—and
charges 8 cents a page.”).
15. New York, New Jersey, California, Arkansas, Tennessee, and Wyoming, among oth-
ers, have all recently or are currently in the process of evaluating logistical, accessibility,
and privacy concerns in the move from paper to online court records. See, e.g., BARRY T.
ALBIN, SUPREME COURT OF N.J., REPORT OF THE SUPREME COURT SPECIAL COMMITTEE ON
PUBLIC ACCESS TO COURT RECORDS i–iii (2007) [hereinafter REPORT ON PUBLIC ACCESS],
available at http://www.judiciary.state.nj.us/publicaccess/publicaccess.pdf (explaining that
“the Committee debated the delicate balance between the public’s general right to know
and the individual’s limited right of privacy within our court system, and [considered] how
placing court records on the Internet will alter exponentially the calculus between those
competing rights,” and outlining the Committee’s recommendations).
16. In New Jersey, for example, the record on appeal need only contain “all papers on
file in the court or courts or agencies below, with all entries as to matters made on the
2012] ONLINE COURT RECORDS 777

enormous variability across jurisdictions, we used the New Jersey


courts as a reference point for honing our ideas about the transition
from paper to electronic records. New Jersey presented a particularly
rich environment because it is in the process of transforming to digi-
tal court filing and record keeping, and has facilitated deliberation
over the ramifications of this transformation, including for privacy. 17
The conclusion of our inquiry is that courts have an obligation to
rewrite rules governing the creation of, and access to, public court
records in light of substantive changes that online access augurs. 18 Al-
though it lies outside the scope of this Article to specify these rules,
we aim to trace a line of analysis toward developing such policy and
regulation, whether through case law or administrative rules. Our
analysis integrates the diverse disciplinary expertise of its authors: tra-
ditional legal sources have informed our grasp of the status quo; 19 a
conception of a right to privacy, drawn from the philosophical theory
of contextual integrity, has guided our evaluation of access policies; 20
and a grasp of underlying technical capacities of information retrieval
systems has shaped the comparative study of information flows, bring-
ing rigor to and generalizing prior work on contextual integrity. 21
How to read this article: What follows is a product of distinct multi-
disciplinary perspectives—legal, philosophical, and technical. Be-
cause each has influenced the others and has shaped our findings and
recommendations, we chose to present an amalgamation of all three
even though the resulting paper does not embody the traditional
form of any one of them. Readers who prefer to focus on arguments
relevant to their particular disciplines may wish to read only those
parts that pique their interest and skip those that are less pertinent. A

records of such courts and agencies, the stenographic transcript or statement of the pro-
ceedings therein, and all papers filed with or entries made on the records of the appellate
court.” N.J. R. 2:5-4(a).
17. See REPORT ON PUBLIC ACCESS, supra note 15, at 9 (discussing the need to balance
the tradition of public access to records with the concern for privacy, and noting that a
comprehensive listing of exceptions to public access is necessary to help reach that bal-
ance).
18. See infra Part V (suggesting that if the rules governing online access to personal in-
formation do not change, then the role of the courts may diminish and parties might
choose other avenues to settle disputes).
19. See infra Part II.B.1 (noting the common law right of access and referencing the
constitutional and common law bases for making court documents open to the public).
20. See infra Part III (discussing the theory of contextual integrity).
21. See infra Part IV.A–B (discussing a model information retrieval system and detailing
the results of actual searches using PACER and Google Scholar); see generally HELEN
NISSENBAUM, PRIVACY IN CONTEXT: TECHNOLOGY, POLICY, AND THE INTEGRITY OF SOCIAL
LIFE (2010) [hereinafter NISSENBAUM, PRIVACY IN CONTEXT] (setting out the theory and
framework of contextual integrity).
778 MARYLAND LAW REVIEW [Vol. 71:772

quick overview will help in this determination. Part II draws mostly on


legal discussions. In particular, Part II.A presents an integral defini-
tion of the term “court record,” Part II.B discusses the sources of a le-
gal right of access to court records, and Part II.C introduces key me-
chanisms by which courts determine whether to grant requests to seal
or redact information contained in their records. Part III provides a
brief philosophical overview of contextual integrity—a theory of in-
formation privacy that guides the inquiry in Part IV and analysis in
Part V. Part IV develops a model of information retrieval for court
records and uses it to compare typical or expected information flows
for in-person access to court records at the courthouse with online
access to court records via PACER and Google Scholar. Focusing on
the transformation from in-person to online access, Part V presents a
comparative normative analysis of information flows, drawing togeth-
er threads from Parts II, III, and IV. The Article concludes in Part VI
with general recommendations for addressing privacy concerns.

II. COURT RECORDS AND THE PRESUMPTION OF ACCESS

A. Definitions
Court records are a subset of government records to which the
public has access. Because the term “court records” is imprecise, we
take this opportunity to explain how we will use this term throughout
the paper. Consider a plaintiff who brings a lawsuit against his boss
for employment discrimination. 22 In order to bring this civil claim,
the employee will need to fill out a complaint form, which details his
claims against his boss, the defendant, and file it with the court. He
must also serve the boss with a summons to appear in court. In reply,
his boss will file an answer to the complaint, responding to the em-
ployee’s allegations and possibly bringing counterclaims of her own.
Through a procedure known as pre-trial discovery, each party’s attor-
ney (if they are both represented) will then request documents and
other evidence from the opposing party, and possibly from relevant
third parties. Before the trial has even begun, each side will likely
have amassed hundreds or even thousands of documents full of in-
formation about the people, places, and events involved in the dis-
pute. The parties will take depositions, which will be recorded or

22. For an explanation of the application of the Federal Rules of Civil Procedure,
which would govern such an action in the federal courts, see generally A.J. STEPHANI &
GLEN WEISSENBERGER, FEDERAL CIVIL PROCEDURE LITIGATION MANUAL (Matthew Bender
ed., 3d ed. 2011). For state procedural rules specifically applicable to this New Jersey case
study, see N.J. R. 4:1–4:101.
2012] ONLINE COURT RECORDS 779

transcribed. Each side will file a variety of additional motions, often


with exhibits and declarations attached. At this point the court may
dismiss the case by summary judgment, the parties may settle out of
court, or the case may go to trial. 23 If the case goes to trial, an even
larger case file will be created that will likely include court transcripts
in writing, audio, or video form, or a combination of the three; 24 wit-
ness testimony; evidence and exhibits; information about jury selec-
tion; 25 additional motions; and, ultimately, a judicial opinion setting
out the final decision in the case.

23. Less than 2 percent of federal civil cases go to trial. ROBERT P. BURNS, THE DEATH
OF THE AMERICAN TRIAL 2 (2009). The pattern is similar for state cases. Id.
24. Sworn officers of the court create transcripts. There are guidelines about what is
and is not included in the transcript, and a delay period before the transcript is placed on
PACER. Parties are notified in advance that the transcript is going “live” and can move to
not have it published or make specific redaction requests. See, e.g., U.S. DIST. COURT DIST.
OF S.C., TRANSCRIPT FILING INSTRUCTIONS FOR ATTORNEYS (2011), available at
http://www.scd.uscourts.gov/CMECF/DOCS/Transcript_Filing_Instructions_Attorneys.p
df (describing how a transcript is made available through PACER after a ninety-day re-
stricted review period, a Notice of Electronic Filing, and the opportunity to redact certain
personal identifiers).
25. In New Jersey, courts do not release a list of juror names at any point during or af-
ter a trial, but names may nevertheless be available in the voir dire transcript. Telephone
Interview with Gina Fe’ Whittaker, Jury Manager, Burlington County Superior Court (Apr.
26, 2011). In high-profile cases, juror names are sometimes kept secret as a precautionary
measure. See KIMBERLEY KEYES, REPORTERS COMM. FOR FREEDOM OF THE PRESS, SECRET
JUSTICE: SECRET JURIES (2005), available at http://www.rcfp.org/secret-justice-secret-
juries/secret-juries (noting that in cases involving terrorism or organized crime, a judge
might choose to use an anonymous jury to protect the jurors). Other state courts, includ-
ing Ohio and Michigan, have found a qualified First Amendment right of access to juror
names and addresses. See State ex rel. Beacon Journal Publ’g Co. v. Bond, 781 N.E.2d 180,
194 (Ohio 2002) (“[W]e hold that the First Amendment qualified right of access extends
to juror names and addresses, thereby creating a presumption of openness that may be
overcome ‘only by an overriding interest . . . .’”) (citation omitted)); In re Disclosure of
Juror Names and Addresses, 592 N.W.2d 798, 809 (Mich. Ct. App. 1999) (“We therefore
hold that the press has a qualified right of postverdict access to juror names and addresses,
subject to the trial court’s discretion to fashion an order that takes into account the com-
peting interest of juror safety . . . .”). The Ohio Supreme Court has also recognized a qual-
ified First Amendment right to access prospective juror questionnaires, which are used to
determine whether potential jurors are suitable for service. Beacon Journal, 781 N.E.2d at
188. In federal courts, however, “documents containing identifying information about
jurors or potential jurors are no longer included in the public case file and are unavailable
to the public, either electronically or at the courthouse.” KEYES, supra. In March 2004, the
Judicial Conference of the United States, the principal policy-making body of the federal
court system, formally adopted this restriction as part of its guidelines about public access
to electronic criminal case files. See JUDICIAL CONFERENCE OF THE UNITED STATES, REPORT
OF THE PROCEEDINGS OF THE JUDICIAL CONFERENCE OF THE UNITED STATES 10 (2004) [he-
reinafter PROCEEDINGS OF THE JUDICIAL CONFERENCE 2004], available at
http://www.uscourts.gov/FederalCourts/JudicialConference/Proceedings/Proceedings.as
px?doc=/uscourts/FederalCourts/judconf/proceedings/2004-03.pdf (adopting a policy
that required, in part, the redaction of personal data identifiers from all court documents
that could be accessed by the public). But see Op. Fla. Att’y Gen. AGO 2005-61, available at
780 MARYLAND LAW REVIEW [Vol. 71:772

Once all the relevant claims have been resolved, whether on


summary judgment or at trial (settlement agreements are typically not
available for public inspection), 26 some number of the above-listed
documents and files will become a permanent part of the court
record27 for this case, for example by being introduced into evidence
as exhibits or attached to motions or other papers filed with the court.
Depending on a number of factors, every document and media file
created during this process may become part of the final court
record, or only certain documents may be entered into the record
while other information-rich documents are excluded. 28 We will dis-
cuss some of the privacy-related justifications for keeping a file out of
the final court record below, 29 but other reasons may be at play as
well, including those related to ease of storage, 30 file or software com-
patibility, 31 or simply local rule or custom. 32 Of course, not all docu-
ments produced during the pre-trial discovery period will ultimately
become part of the court record, but a fair number may be included,
particularly those that the parties rely on in their moving papers.
In this Article, we will use the term “case file” to refer to the sum
total of documents, media files, and exhibits that are produced and
collected by the parties and/or the court as a case makes its way
through the system. We will use “court record” to refer to the subset
of these documents that remain after a case has been resolved and

http://myfloridalegal.com/ago.nsf/Opinions/4A410DA4389543CC852570C000750D51
(stating that juror names and addresses are not exempt from public disclosure pursuant to
a Florida state public records statute). For a general discussion of the risks and benefits of
juror anonymity, see Christopher Keleher, The Repercussions of Anonymous Juries, 44 U.S.F. L.
REV. 531, 547–59 (2010) (discussing how juror anonymity affects such elements as juror
privacy and safety, voir dire, summation, and right of access to trial proceedings).
26. See Sharona Hoffman, Settling the Matter: Does Title I of the ADA Work?, 59 ALA. L. REV.
305, 313 (2008) (noting that “settlement amounts are generally not recorded in publicly
available court documents”).
27. This term should be distinguished from the official Appellate Record (or the
Record on Appeal)—the record an appellate court has access to and reviews in an ap-
peal—which may or may not be identical to the publicly available record at the trial court.
See, e.g., supra note 16 and accompanying text.
28. See infra Part II.B.2.c (discussing how certain documents, or parts of documents,
may be let in or kept out of court records by various means, including state rules, sealing
of records, or redaction of personal information).
29. See infra Part II.B.2.c.
30. See infra text accompanying notes 111–113.
31. Cf. Robert A. Guy Jr., A Model Protocol for Electronic Filing: Best Practices for Law Firms
Making the Transition to Case Management/Electronic Case Files, 53 FED. LAW. 38, 45, Aug. 2006
(noting some potential errors, including software errors, that may occur when an attorney
is filing court documents electronically into an electronic case management system).
32. See infra Part IV.C.2.b-c.
2012] ONLINE COURT RECORDS 781

become part of the permanent, public record. 33 If a court chooses to


allow remote electronic access to its records, the entire court record
may be placed online, or only a subset or summary of the record may
be available. As we discuss at length in Part IV, online access may be
limited by logistic constraints, may be available only to certain catego-
ries of persons, may be located behind a paywall, or some combina-
tion of the three.
As we have defined it, a case file may include thousands of doc-
uments full of information about the parties and the dispute. It may
contain motions; pleadings; briefs; attachments; dockets; transcripts
in textual, audio, 34 or video form; exhibits entered into evidence; and
records and responses to interrogatories and requests for admission
produced during pre-trial discovery. Each and every form filled out
by the parties, their lawyers, or by related third parties (witnesses, ju-
rors, etc.) potentially contains vast amounts of personal data includ-
ing home or school addresses, places of employment, birthdates, and,
in many cases, Social Security numbers. 35 While much of this infor-
mation may appear innocuous and uninteresting to anyone but the
parties themselves, some of it may be quite revealing or even embar-
rassing for the individuals and organizations involved in a dispute. 36
Exhibits entered into evidence may include bank statements, medical
records, psychological evaluations, personal and business emails, and
other intimate details about each party and their interactions with
one another or with others. All or a portion of this file may become
part of the public record, available in paper form at the courthouse

33. It bears noting that while we are making this analytical distinction—between the
case file and the court record—it is not necessarily accurate in practice. Documents in the
case file may become public at the time of trial, for example, complicating our categoriza-
tion. Thus, we make this distinction here only for analytical purposes.
34. See, e.g., JUDICIAL CONFERENCE OF THE UNITED STATES, REPORT OF THE
PROCEEDINGS OF THE JUDICIAL CONFERENCE OF THE UNITED STATES 9–10 (2010) [hereinaf-
ter PROCEEDINGS OF THE JUDICIAL CONFERENCE 2010], available at
http://www.uscourts.gov/FederalCourts/JudicialConference/Proceedings/Proceedings.as
px?doc=/uscourts/FederalCourts/judconf/proceedings/2010-03.pdf (stating that the
Judicial Conference adopted a recommendation that digital audio files be available to the
public through the PACER system, especially because the availability of such recordings
improved public access).
35. For example, a person filing a complaint with the New Jersey courts is required to
provide the following information: name, address, phone number, name of defendant,
defendant’s address, summary of issue that prompted filing of complaint, summary of
harm caused by defendant, date, and signature. See N.J. Civil Compl. Form, available at
http://www.judiciary.state.nj.us/civil/forms/11210_civil_action_complaint.pdf.
36. See Marder, supra note 2, at 445–47 (noting that the personal disclosures that occur
in litigation may include the names of rape victims, evidence of marital infidelity, personal
assets and debts, and other information that the parties involved wish to keep private).
782 MARYLAND LAW REVIEW [Vol. 71:772

and, in some states and in the federal court system, 37 via remote
access in whole or in part on the Internet.
Even if the information in the case file does not itself immediate-
ly appear revealing, when combined with other publicly available da-
ta, such as phone and property records, it may provide ample infor-
mation for identity thieves. 38 Especially (but not exclusively) if the
dispute took place several years ago, before the widespread use of So-
cial Security numbers in identity theft became a major concern, doc-
uments may be riddled with this information. 39 In most states, the
burden is on the lawyers and their clients to redact sensitive informa-
tion like Social Security numbers before filing a document with the
court. 40 Even the most well intentioned counsel, however, may let a
few items slip, particularly in files cataloged by hand which may in-
clude written Social Security numbers in the upper right corner of the
file. 41 When these documents are scanned into portable document
format (“PDF”), hand-written Social Security numbers will be nearly
impossible to locate and remove through any software that a court
might use to detect and black out sensitive information. 42 Signatures

37. For an in-depth discussion of the PACER system, see infra Part IV.
38. See Marder, supra note 2, at 447 (noting that personal information in case files can
lead to other personal information available elsewhere on the Internet, which can then be
compiled and used to commit identity theft).
39. For an extensive discussion of the use of Social Security numbers in certain docu-
ments, the associated risk of identity theft, and legislation and court cases addressing So-
cial Security numbers, see Social Security Numbers, ELECTRONIC PRIVACY INFORMATION
CENTER, http://epic.org/privacy/ssn/ (last visited Mar. 5, 2012).
40. New Jersey, for example, “requires the parties to redact any confidential personal
identifiers from documents submitted to the court, but . . . fails to provide an enforcement
mechanism to punish a party’s lack of compliance.” Kristin M. Makar, Comment, Taming
Technology in the Context of the Public Access Doctrine: New Jersey’s Amended Rule 1:38, 41 SETON
HALL L. REV 1071, 1073 (2011).
41. See Andy Opsahl, Privacy: Agencies Struggle to Redact Personal Data from Online Public
Documents, GOV’T TECH. (July 8, 2008), http://www.govtech.com/gt/375540 (“These doc-
uments frequently contain Social Security numbers (SSNs), mothers’ maiden names, sig-
natures, minors’ names and other red meat for identity thieves and stalkers. . . . To redact
SSNs, states use software to black out the section of the document where the number ap-
pears. Redaction software vendors typically guarantee 98 percent accuracy, but SSNs tend
to crop up everywhere in documents, increasing the difficulty of redacting them. For ex-
ample, some lenders required borrowers to write their SSNs beneath their signatures.
Other lenders organized the documents by SSN within their office filing systems. To do
that, they wrote the SSNs on the upper right corners of the documents.”); see also Jacob
Ogles, Court Documents Not Fit For Web?, WIRED.COM (Nov. 23, 2004),
http://www.wired.com/politics/security/news/2004/11/65703 (“Though most states re-
quire information like Social Security numbers or dates of birth to be concealed before
documents are posted on a web server, it is impractical and sometimes impossible for
clerks to catch every instance.”).
42. See Timothy B. Lee, Studying the Frequency of Redaction Failures in PACER, FREEDOM
TO TINKER (May 25, 2011, 1:52 PM), https://freedom-to-tinker.com/blog/tblee/studying-
2012] ONLINE COURT RECORDS 783

on forms may be scanned as well, and if these scans are ultimately


placed on the web, their presence further increases the risk of identity
theft. 43
Case files may contain personally identifiable or otherwise reveal-
ing information not only about the parties in a dispute, but also fami-
ly members, colleagues, witnesses, jurors, and even victims or wronged
parties in criminal or civil cases, respectively. 44 Laws regulating when
and how such information may be disclosed, particularly with respect
to victim and juror information, vary by state. For example, the web-
site of a Massachusetts victims’ rights group warns sexual assault vic-
tims that while their names will be blacked out in all police reports
and court records, they must file a formal request with the judge to
have their address, telephone number, place of employment, or
school location redacted. 45
While the focus of this Article is on court records of civil cases, it
bears noting that when an individual is arrested or indicted, this in-
formation will most likely become part of the publicly available court
record even if the charges are ultimately dismissed. 46 Thus, even in cases
where charges were wrongfully brought—a case of mistaken identity,
perhaps, or simply a misunderstanding—the record of that individu-
al’s history in the criminal justice system will remain. Rarely will this

frequency-redaction-failures-pacer; Timothy B. Lee, What Gets Redacted in Pacer, FREEDOM


TO TINKER (June 16, 2011, 1:48 PM), https://freedom-to-tinker.com/blog/tblee/what-
gets-redacted-pacer.
43. Opsahl, supra note 41.
44. See infra Part II.C (discussing the privacy concerns of unrepresented third parties).
45. The website reads:
In Massachusetts, if you make a report to the police about sexual assault, [Massa-
chusetts law] makes sexual assault reports to the police unavailable to the gener-
al public. They are kept in separate, confidential files. However, if your case be-
comes active in the criminal courts, your police report will be made public.
Another Massachusetts law . . . requires that your name should automatically be
blacked-out (“redacted”) in all public police records and court documents. This
means that your name should not be released to the public. You also have the
additional option to ask the Judge to keep your address, telephone number,
place of employment, and/or location of your school private (“impounded.”). If
the Judge agrees, this personal information will be redacted from the record and
not be stated in open court. Please note, in order to do this, you must request
this from the Judge as soon as possible.
Privacy, VICTIM RIGHTS LAW CENTER, http://www.victimrights.org/node/109 (last visited
Mar. 5, 2012).
46. See, e.g., COMMONWEALTH OF MASSACHUSETTS, DISTRICT COURT DEPARTMENT OF
THE TRIAL COURT, A GUIDE TO PUBLIC ACCESS, SEALING & EXPUNGEMENT OF DISTRICT
COURT RECORDS 11–13 (2010), available at http://www.mass.gov/courts/courtsandjudges
/courts/districtcourt/pubaccesscourtrecords.pdf (listing scenarios in criminal cases where
information remains public despite the case being dismissed, the prosecution being aban-
doned, or the verdict being not guilty).
784 MARYLAND LAW REVIEW [Vol. 71:772

record note that the charges were dismissed, or that the individual
was found to be innocent. Instead, the record will likely end with the
arrest or arraignment, providing no further details. A potential em-
ployer can easily access this information and may reasonably—though
mistakenly—assume that the applicant has a criminal record. 47 This is
particularly disconcerting given that low-income individuals and
people of color are considerably more likely to be wrongfully accused
or arrested. 48

B. Legal Basis for the Right of Access

1. Sources of the Right of Access


Under the Freedom of Information Act (“FOIA”), passed in
1966, the majority of government records in the United States are
open to public inspection. 49 If you are curious what the Food and
Drug Administration or the Department of Homeland Security are up
to, a FOIA request can provide that information (including that
which is not already made available on agencies’ respective web-

47. This type of unfair discrimination is not limited to employment opportunities. In


Minnesota, “banks used court records to pore through financial and other records of mi-
nority applicants and then decided to deny home loans to some of those individuals, ac-
cording to a Minnesota court advisory committee report issued in September.” Ogles, su-
pra note 41.
48. This assumes that white and African-American individuals commit crimes at equiva-
lent rates. See Vincent Schiraldi & Jason Ziedenberg, Race and Incarceration in Maryland,
JUST. POL’Y INST. 6 (Oct. 23, 2003), http://www.justicepolicy.org/uploads
/justicepolicy/documents/03-10_rep_mdraceincarceration_ac-md-rd.pdf (stating that the
likelihood of being incarcerated is seven to eight times greater for African-Americans than
for whites). Harvey Grossman, the Legal Director for the ACLU of Illinois, stated:
In looking at this data [from the Illinois Traffic Stops Study], it is reasonable to
conclude that for police in Illinois, a driver’s race is a proxy for suspicion in de-
ciding whether to request a consent search. . . . It is clear that police view drivers
of color with far more suspicion than their white counterparts.
ROGER BALDWIN FOUNDATION OF ACLU, ANNUAL REPORT 2008-2009: BUILDING AN
AMERICA WE CAN BE PROUD OF 9, available at http://www.aclu-il.org/wp-
content/uploads/2011/01/Annual-Report-2009-webversion.pdf. Racial profiling is not
limited to traffic stops. The Justice Policy Institute found that predominantly African-
American neighborhoods are targeted more often for violations of drug laws. Schiraldi &
Ziedenberg, supra, at 16. Despite the fact that “African Americans and whites use drugs at
a comparable rate, [African Americans] represent 68% of those arrested for drug offenses,
and 90% of those incarcerated for drug offenses” nationwide. Ngozi Caleb Kamalu et al.,
Racial Disparities in Sentencing: Implications for the Criminal Justice System and the African Ameri-
can Community, 4 AFR. J. CRIMINOLOGY & JUST. STUD. 1, 15 (2010).
49. Freedom of Information Act, 5 U.S.C. § 552 (2006). Most states have an analogous
state statute providing access to state government records. See generally ANNE WELLS
BRANSCOMB, WHO OWNS INFORMATION? FROM PRIVACY TO PUBLIC ACCESS 8 (1994) (pro-
viding an overview of FOIA and privacy legislation in the United States).
2012] ONLINE COURT RECORDS 785

sites). 50 But requests pursuant to this act cannot reach records of the
judicial branch—namely, court records—because FOIA applies by its
terms only to agency records (records of the executive branch). 51
The tradition of providing public access to court records, howev-
er, pre-dates FOIA. 52 Given all the privacy concerns noted above, one
reasonably might ask why the public is given a right to access such
records in the first place. 53 But the right to open courts and their
records is actually as longstanding as our right to the courts and to
justice itself: it is based on the widely held belief that for a justice sys-
tem to function successfully and consistently, it must be accountable
to its citizens. 54 In theory, open access to everything that happens
within this system, from the first filing of a lawsuit or of criminal
charges, to the final hearing for damages or sentencing, provides
journalists and concerned citizens alike with the opportunity to en-
sure that justice is being apportioned fairly and consistently through-
out the country. 55 Judicial openness has two components: the right of
access to trials themselves, and the right of access to judicial docu-
ments for inspection and copying. 56 The right of public access to

50. See generally REPORTERS COMM. FOR FREEDOM OF THE PRESS, FEDERAL OPEN
GOVERNMENT GUIDE (10th ed. 2009), available at http://www.rcfp.org/federal-open-
government-guide (describing the various components of FOIA, including how to make a
request, what agencies are subject to those requests, and how to track those requests).
51. The E-Government Act of 2002 does provide for public access to federal records
across all three branches of government. Under this Act, federal courts are required to
establish and maintain websites with docket information for pending cases and closed cas-
es (up to one year old), electronically filed court documents, judicial opinions in text sear-
chable format “regardless of whether such opinions are to be published in the official
court reporter,” and “[a]ny other information . . . that the court determines useful to the
public.” E-Government Act of 2002, Pub. L. No. 107-347, § 205(a)–(b), 116 Stat. 2910,
2913–14 (codified at 44 U.S.C. § 3501 (2006)). It should be noted that not all electronical-
ly filed court documents are subject to disclosure. See id. at § 205(c).
52. See, e.g., UNITED STATES DEP’T OF JUSTICE, ATTORNEY GENERAL’S MANUAL ON THE
ADMINISTRATIVE PROCEDURE ACT 24 (1947) (discussing the Administrative Procedure Act,
which pre-dated FOIA, and when records could be made available under the APA).
53. One may also ask why we do not litigate anonymously. See generally Lior Jacob Stra-
hilevitz, Pseudonymous Litigation, 77 U. CHI. L. REV. 1239 (2010) (providing further discus-
sion of the potential merits of anonymous pseudonymous litigation).
54. See Globe Newspaper Co. v. Superior Court, 457 U.S. 596, 606 (1982) (“Public scru-
tiny of a criminal trial enhances the quality and safeguards the integrity of the factfinding
process, with benefits to both the defendant and to society as a whole.”).
55. See id. (“[P]ublic access to the criminal trial fosters an appearance of fairness, the-
reby heightening public respect for the judicial process.”).
56. See Richmond Newspapers, Inc. v. Virginia, 448 U.S. 555, 580 (1980) (plurality opi-
nion) (“We hold that the right to attend criminal trials is implicit in the guarantees of the
First Amendment . . . .” (footnote omitted)); see also United States v. Gotti, 322 F. Supp. 2d
230, 239 (E.D.N.Y. 2004) (discussing both the right of access to judicial proceedings and
right of access to judicial documents).
786 MARYLAND LAW REVIEW [Vol. 71:772

judicial documents predates the Constitution and may be traced as far


back as English common law, the predecessor to the U.S. legal sys-
tem. 57
While the common-law right of access to judicial documents has
been well recognized, whether such a right is separately rooted in the
Constitution remains somewhat of an open question. 58 In Richmond
Newspapers, Inc. v. Virginia, the Supreme Court held that both the First
and Fourteenth Amendments grant the press and the public the right
to attend criminal trials. 59 But the right to attend judicial proceedings
does not necessarily give one access to any judicial documents used in
or produced by such proceedings. 60 To wit, in Zenith Radio Corp. v.
Matsushita Electric Industrial Co. a federal district court held that while
the Supreme Court has recognized First and Fourteenth Amendment
rights of access to criminal trials themselves, 61 the second component
of judicial openness—access to all types of judicial records for inspec-
tion and copying—is not constitutionally protected. 62 Thus, while the

57. United States v. Amodeo, 44 F.3d 141, 145 (2d Cir. 1995) (citing Leucadia, Inc. v.
Applied Extrusion Techs., Inc., 998 F.2d 157, 161 (3d Cir. 1993)); see also United States v.
Criden, 648 F.2d 814, 819 (3d Cir. 1981) (further supporting that the right to inspect pub-
lic records predates the Constitution); Gotti, 322 F. Supp. 2d at 239 (“The common law
right of access to judicial documents under American jurisprudence traces its origin to the
general English common law right of access to public records, but has a broader reach.”).
58. The Supreme Court has laid out a two-part test for determining whether the public
has a First Amendment right of access to judicial proceedings: (1) Has the proceeding his-
torically been open? (2) Does the right of access play an essential role in the functioning
of the judicial process and the government as a whole? See In re Reporters Comm. for
Freedom of the Press, 773 F.2d 1325, 1331–32 (D.C. Cir. 1985) (“Apparently, both these
questions must be answered affirmatively before a constitutional requirement of access can
be imposed.”); see also Globe Newspaper Co., 457 U.S. at 606 (noting that the right of access
to court records is based on “constitutional stature,” but is “not absolute”); Richmond News-
papers, 448 U.S. at 569–71 (discussing how historical evidence supports that the right to
open access to trials stems from common law).
59. Richmond Newspapers, 448 U.S. at 580.
60. Zenith Radio Corp. v. Matsushita Elec. Indus. Co., 529 F. Supp. 866, 897 (E.D. Pa.
1981).
61. Id. at 895 (citing Richmond Newspapers, 448 U.S. 555).
62. Id. at 908 (“With respect to the question whether the common law right to inspect
and copy has a constitutional dimension, we conclude that it does not.”); see also Brown &
Williamson Tobacco Corp. v. FTC, 710 F.2d 1165, 1176–81 (6th Cir. 1983) (holding that
the First Amendment limits judicial discretion to seal court documents but also stating that
the right of access is still not absolute); Gotti, 322 F. Supp. 2d at 243–50 (noting a qualified
First Amendment right of access to both judicial proceedings and judicial documents).
But see Republican Co. v. Appeals Court, 812 N.E.2d 887, 892 (Mass. 2004) (“The exercise
of the power to restrict access, however, must recognize that impoundment is always the
exception to the rule, and the power to deny public access to judicial records is to be
‘strictly construed in favor of the general principle of publicity.’” (citation omitted)).
2012] ONLINE COURT RECORDS 787

right of access remains available under common law, the First


Amendment may not add any additional protection. 63

2. Contours of the Right of Access


In 1978, the Supreme Court began to clearly articulate the con-
tours of the common-law right of access to court records and the justi-
fications for this right. 64 Federal and state courts and legislatures have
picked up this project as well, outlining the precise shape of this right
and enshrining it in both common and state statutory law. 65
As part of a broader discussion about the purposes of public
access, federal and state courts have addressed such questions as: Who
is entitled to access judicial documents? 66 What types of documents
are included in this right? 67 And when should access be allowed or
denied based on concerns about secrecy or control over informa-
tion? 68 While the presumption of access to court records is longstand-
ing, it has never been interpreted to allow access to everything in the
case file. 69 These restrictions on access trickle down from state and
federal appellate courts to the local courthouses themselves, where
state and local law, custom, and in some cases simply the whims of
court clerks determine which information in the court record will ac-
tually be made available to the public, and how.
Within the broad mandate to provide access, courts are given
some leeway in terms of how they implement the public’s right to in-
spect and copy judicial records. 70 In the following section, we provide
a basic outline of the restrictions on access to court records as well as
the balancing tests judges employ to determine when and why some
court records, in whole or in part, will not be available for public in-
spection. At each point, it is important to remember that the law re-
mains unclear at the level of individual access in part because the sta-

63. Zenith Radio Corp., 529 F. Supp. at 913 (concluding “that a constitutional right of
access to judicial records does not exist”).
64. Nixon v. Warner Commc’ns, Inc., 435 U.S. 589, 597, 607 (1978).
65. Daniel J. Solove, Access and Aggregation: Public Records, Privacy and the Constitution, 86
MINN. L. REV. 1137, 1158–63 (2002).
66. See infra Part II.B.2.a.
67. See infra Part II.B.2.c.
68. See infra Part II.B.2.c.
69. Zenith Radio Corp. v. Matsushita Elec. Indus. Co., 529 F. Supp. 866, 897–901 (E.D.
Pa. 1981).
70. For an extensive discussion of the legal basis of the right to access court records, as
well as the specific contours of this right in each U.S. state, see REPORTERS COMM. FOR
FREEDOM OF THE PRESS, ELECTRONIC ACCESS TO COURT RECORDS, SPRING 2007, available at
http://www.rcfp.org/rcfp/orders/docs/EACR.pdf.
788 MARYLAND LAW REVIEW [Vol. 71:772

tutory language is not itself precise, and in part because many of the
decisions about access are made on an ad hoc basis by individual
court clerks or other employees. 71

a. Who May Access: The Actors


We all have the image in our minds of a concerned citizen going
down to the local courthouse to page through old court records, per-
haps looking for information on land title or family history. But the
details of this scenario may be a bit fuzzy in our heads. To which
records does she have access? May she remove them from the court-
house? Make copies of them at a local terminal? Does she need to
prove that she has some connection to these records, some valid in-
terest in acquiring them? Is there a restriction on what she may do
with the information in the records once she obtains it?
Generally speaking, an individual need only show a “legitimate
interest” in the public record requested in order to gain access. 72 As
early as 1908, the Supreme Court of Alabama held that the clerk of
the court was compelled by statute to provide copies of papers held by
the court to any person requesting them, and had no power to de-
termine whether the requester should be granted the copies sought. 73
In 1995, the New Jersey Supreme Court determined that even a pure-
ly for-profit interest in judicial records counts as “legitimate,” and
therefore allowed a seller of municipal tax-assessment data and a
commercial real estate appraiser to obtain copies of computerized
records of municipal tax assessments for use in their business enter-
prises. 74

71. See infra Part II.B.2.c & note 107. This is not to suggest that all courts do not have
clear access policies. Many do, and many others are in the process of constructing them.
72. See, e.g., Nixon v. Warner Commc’ns, Inc., 435 U.S. 589, 597–98 (1978) (“American
decisions generally do not condition enforcement of this right on a proprietary interest in
the document or upon a need for it as evidence in a lawsuit. The interest necessary to
support the issuance of a writ compelling access has been found, for example, in the citi-
zen’s desire to keep a watchful eye on the workings of public agencies . . . and in a news-
paper publisher’s intention to publish information concerning the operation of govern-
ment.” (citation omitted)); Irval Realty Inc. v. Bd. of Pub. Util. Comm’rs, 294 A.2d 425,
428 (N.J. 1972) (a citizen requesting records must “show an interest in the subject matter
of the material he sought to scrutinize. Such interest need not have been purely personal.
As one citizen or taxpayer out of many, concerned with a public problem or issue, he
might demand and be accorded access to public records bearing upon the problem, even
though his individual interest may have been slight.”).
73. Jackson v. Mobley, 47 So. 590, 593 (Ala. 1908). But see Nixon, 435 U.S. at 598
(“Every court has supervisory power over its own records and files, and access has been de-
nied where court files might have become a vehicle for improper purposes.”).
74. Higg-A-Rella, Inc. v. County of Essex, 660 A.2d 1163, 1169 (N.J. 1995). But see Bur-
nett v. County of Bergen, 968 A.2d 1151, 1164–66 (N.J. 2009) (stating that while a com-
2012] ONLINE COURT RECORDS 789

b. How Records Can Be Accessed at the Courthouse:


Transmission Principles
Whether a record-seeker needs to show identification, pass
through a metal detector, or sign in to access records depends entire-
ly on the particular court she is visiting. 75 At the New Jersey Supreme
Court, she would need to provide identification and receive a name
badge with her photo on it in order to enter the courthouse; if she
has already visited before, her name will be stored in the courthouse
security database, and her original name badge will be printed upon
providing identification. To access the records maintained at the Su-
preme Court, she will need to sign in once she reaches the records
room.
Most courts provide copies of records for a nominal fee. The
New Jersey state courts will provide records “only in the form in which
they are maintained or indexed by the Judiciary” and state in their
official rules that “[r]equests by private individuals or entities for pro-
gramming, searching, or compilation of records in a form other than
as used for the Judiciary’s purposes will not be granted.” 76 Thus, in
New Jersey our hypothetical record-seeker could, for a small fee, 77 ob-
tain a copy of any publicly available court record in the format in
which it is kept. She could not, however, request that the clerk pro-
vide her with a compilation of all records involving civil suits with
damage awards greater than $50,000, or all records from cases involv-
ing noise ordinances. 78 Once she obtains a copy of the publicly avail-
able records, she is free to use them however she chooses, provided
that she does not violate any state or federal laws. 79

mercial entity has the same rights to judicial records as anyone else, “the compilation of
otherwise hard-to-obtain information alters the privacy interest implicated by disclosure of
that information” (citation and internal quotation marks omitted); see also infra Part V
(discussing balancing tests and electronic access).
75. See infra Part IV (comparing in-person access to court records in New Jersey to elec-
tronic access).
76. N.J. R. 1:38-13.
77. Id. New Jersey Courts Rule 1:38-9 states that “[t]he Supreme Court shall establish a
schedule of fees for copies of records.” Id. at 1:38-9.
78. See id. at 1:38-13 (stating that records are only available in the form in which they
are “maintained or indexed” by the judiciary, which implies that compilations of records
from multiple cases of the same type are not available for public access).
79. Certain uses of information could, for example, result in the commission of a pri-
vacy tort. See Cox Broad. Corp. v. Cohn, 420 U.S. 469, 474, 487 (1975) (discussing the in-
dividual right of privacy and court actions associated with that right with respect to infor-
mation contained in publicly accessible records).
790 MARYLAND LAW REVIEW [Vol. 71:772

c. What May Be Accessed: Information Type


Explicit Inclusions: Each state court may indicate via statute what
material is and is not included in the publicly accessible court
record. 80 New Jersey Courts Rule 1:38, “Public Access to Court
Records and Administrative Records,” provides a definition of the
term “court record” based on what it does and does not include. As
stated at the beginning of this rule, New Jersey court records include:
(1) any information maintained by a court in any form in
connection with a case or judicial proceeding, including but
not limited to pleadings, motions, briefs and their respective
attachments, evidentiary exhibits, indices, calendars, and
dockets;
(2) any order, judgment, opinion, or decree related to a
judicial proceeding;
(3) any official transcript or recording of a public judicial
proceeding, in any form;
(4) any information in a computerized case management
system created or prepared by the court in connection with
a case or judicial proceeding;
(5) any record made or maintained by a Surrogate as a judi-
cial officer. 81
The mandate for federal courts is considerably more vague.
Federal Rule of Civil Procedure 79 requires the clerk of each federal
court to “keep a record known as the ‘civil docket,’” which includes
“papers filed with the clerk” as well as “appearances, orders, verdicts,
and judgments,” and “any other records required by the Director of
the Administrative Office of the United States Courts with the ap-
proval of the Judicial Conference of the United States.” 82 This rule is
implemented via the PACER system, and discussed in Part IV.
Explicit Exclusions: Some information is excluded from the court
record by definition. According to Rule 1:38-2(b), New Jersey court
records do not include “information gathered, maintained or stored
by a governmental agency or . . . unfiled discovery materials.” 83 The
rule also lists specific types of records that are excluded from public
access, including, but not limited to, records pertaining to: juvenile

80. Ex parte Capital U-Drive-It, Inc., 630 S.E.2d 464, 469 (S.C. 2006) (discussing how
public access “[r]estrictions may be based on a statute or the court’s inherent power to
control its own records and supervise the functioning of the judicial system”).
81. N.J. R. 1:38-2(a).
82. FED. R. CIV. P. 79.
83. N.J. R. 1:38-2(b).
2012] ONLINE COURT RECORDS 791

delinquency, victims of abuse or sexual assault, child placement, and


state mandated drug treatment programs. 84 Other states, however,
such as Massachusetts, do not provide for the automatic exclusion of
personally identifying information of victims of sexual assault, so the
burden is on individuals concerned about their privacy or the privacy
of their family members to be aware of state law and, if necessary, to
request that sensitive records be excluded with a showing of good
cause. 85
Some records that would otherwise be available to the public may
instead be “sealed.” Even before a document ever becomes part of
the court record, a judge has the option to seal that document. 86
Thus, a judge may seal entire case files under certain circumstances 87
or only certain documents in the file, such as discovery materials 88 or
financial records. 89 A judge may decide to seal a record on her own, 90
at the request of one party, 91 or at the request of both parties. 92 Even
if a record is not sealed, some information may be redacted before fil-
ing. 93 Each party’s lawyer is responsible for redacting sensitive ma-
terial, such as Social Security and bank account numbers, from her
party’s filings before they become public. 94 In some cases, parties may
request that the judge redact certain otherwise public information

84. Id. at 1:38-3(c)–(d)


85. See supra note 45 (noting that individuals must file formal requests with the judge
in order to have certain information redacted from their record).
86. FED R. CIV. P. 5.2(d).
87. See, e.g., Joy v. North, 692 F.2d 880, 893 (2d Cir. 1982) (finding that adjudications
can be sealed, but only under “exceptional circumstances”).
88. Seattle Times Co. v. Rhinehart, 467 U.S. 20, 37 (1984) (affirming the issuance of a
protective order limited to pretrial civil discovery).
89. See United States v. Hickey, 997 F. Supp. 1206, 1208–09 (N.D. Cal. 1998) (noting
that unsealing defendants’ financial affidavits would result in “substantial and real hazards
of incrimination”).
90. See In re Knight Publ’g Co., 743 F.2d 231, 235 (4th Cir. 1984) (“The trial court . . .
may, in its discretion, seal documents if the public’s right of access is outweighed by com-
peting interests.”).
91. FED. R. CIV. P. 26(c)(1).
92. See Laurie Kratky Doré, Secrecy by Consent: The Use and Limits of Confidentiality in the
Pursuit of Settlement, 74 NOTRE DAME L. REV. 283, 385–86 (1999) (explaining that “litigants
might further condition their compromise on the sealing . . . of particular documents”).
93. See Katie Mulvaney, ACLU Objects to Redaction of Federal Court Records, PROVIDENCE J.
BULL., Dec. 7, 2010 (reporting that “[w]hile all federal courts are required to hide an indi-
vidual’s Social Security number, birth dates and minors’ names from publicly accessible
court records,” the United States District Court for the District of Rhode Island considered
a proposal that would allow parties “to ask the court to have information other than per-
sonal identifiers . . . redacted from court transcripts”).
94. See FED. R. CIV. P. 5.2(a) (stating that “a party . . . making the filing may include
only” certain information in the document).
792 MARYLAND LAW REVIEW [Vol. 71:772

based on privacy interests or concerns about trade secrets. 95 Depend-


ing on the court, federal or state law controls whether a record will be
sealed or redacted. 96
In federal court, the Federal Rules of Civil Procedure require
parties to include only: the last four digits of their Social Security
number, taxpayer identification number, or financial account num-
ber; the year of their birth rather than the entire birth date; and the
initials of any minor identified in the case. 97 When other privacy con-
cerns arise, the court may allow individuals to file certain documents
under seal, but without redaction. 98 “The court may later unseal the
filing or order [the party] to file a redacted version for the public
record.” 99 Alternatively, the court may grant a party’s request that ad-
ditional information be redacted from the record, or be made un-
available via remote access, upon a showing of good cause. 100
Generally speaking, most discovery materials collected by the
parties are excluded from the publicly available record, on the theory
that providing access to them does not further the goal of ensuring
honesty and respect in the judicial system.101 “Discovery is traditional-
ly a private aspect of litigation”—it happens not in a public forum but
behind the scenes and before the trial—therefore, “public scrutiny
would have little value.” 102 Further, providing access to discovery ma-
terials may make an already cumbersome and expensive process even
more so. 103 Of course, some discovery materials are admitted as evi-

95. See Peter A. Winn, Judicial Information Management in an Electronic Age: Old Standards,
New Challenges, 3 FED. CTS. L. REV. 135, 138, 142 (2009) (noting that courts may use redac-
tion to prevent access to trade secrets and “other sensitive information”).
96. See infra Part II.C.
97. FED. R. CIV. P. 5.2(a). But see id. at 5.2(b) (discussing exemptions).
98. Id. at 5.2(d).
99. Id.
100. Id. at 5.2(e). Courts have interpreted “good cause” in a variety of ways. See infra
Part II.C and text accompanying notes 144–159.
101. See Leucadia, Inc. v. Applied Extrusion Techs., Inc., 998 F.2d 157, 164–65 (3d Cir.
1993) (noting that “[t]he public policy implications of” affording a presumptive right of
access to discovery materials “are unclear”); Anderson v. Cryovac, Inc., 805 F.2d 1, 12 (1st
Cir. 1986) (finding that public access to discovery does not “play a significant role in the
administration of justice”).
102. Diane Apa, Common Law Right of Public Access—The Third Circuit Limits Its Expansive
Approach to the Common-Law Right of Public Access to Judicial Records: Leucadia Inc. v. Applied
Extrusion Techs., Inc., 39 VILL. L. REV. 981, 1000 (1994).
103. See Anderson, 805 F.2d at 12 (“Indeed if such cases were to be mandated, the civil
discovery process might actually be made more complicated and burdensome than it al-
ready is.”).
2012] ONLINE COURT RECORDS 793

dence at the trial, at which point the presumption of access to these


materials applies. 104
Inclusions by omission: Even after taking account of procedural
rules or state statutes regarding information that may be sealed, re-
dacted, and excluded from the record, precisely what the general
public is entitled to access in each court record remains somewhat of
an open question. Generally speaking, courts have held that once
something has become “public,” that is, made available for public
viewing or listening in some form, even if only to a portion of the
public, it remains public for the purposes of the court record. 105 For
example, the United States Court of Appeals for the Second Circuit
has stated:
Once the evidence has become known to the members of
the public, including representatives of the press, through
their attendance at a public session of court, it would take
the most extraordinary circumstances to justify restrictions
on the opportunity of those not physically in attendance at
the courtroom to see and hear the evidence, when it is in a
form that readily permits sight and sound reproduction. 106

104. See Littlejohn v. Bic Corp., 851 F.2d 673, 683 (3d Cir. 1988) (holding that the
“right of access [applies] to items that properly remained part of the judicial record, such
as the deposition testimony read into evidence at trial”).
105. See United States v. Martin, 746 F.2d 964, 967–69 (3d Cir. 1984) (finding that
“[t]he public interest can best be vindicated by the release of complete and accurate tran-
scriptions” of audiotapes played for the jury at trial and not simply by reporters’ presence
in the courtroom during the trial to take notes on the recordings). Note that with this
move, courts are vastly expanding the meaning of “public” as it relates to court records. It
may not be feasible or even desirable for the majority of the “public” to attend or other-
wise listen to a trial, yet the record becomes available to everyone, including those who live
far from the courthouse, especially if the record is later placed online. See, e.g., United
States v. Criden, 648 F.2d 814, 815 (3d Cir. 1981) (granting television networks’ request
“‘for permission to copy, for the purpose of broadcasting to the public, those video and
audio tapes admitted into evidence and played to the jury in open court’”); Lynn E. Sud-
beck, Placing Court Records Online: Balancing Judicial Accountability with Public Trust and Con-
fidence: An Analysis of State Court Electronic Access Policies and a Proposal for South Dakota Court
Records, 51 S.D. L. REV. 81, 91 (2006) (noting that a “frequently mentioned benefit” of
electronic access to court records is that it “is appropriate to the needs of South Dakota’s
rural court users, that is, [it] levels the geographic playing field” (citation and internal qu-
otation marks omitted)).
106. In re Nat’l Broad. Co., 635 F.2d 945, 952 (2d Cir. 1980); see also Cox Broad. Corp. v.
Cohn, 420 U.S. 469, 494–95 (1975) (“[T]he prevailing law of invasion of privacy generally
recognizes that the interests in privacy fade when the information involved already appears
on the public record. The conclusion is compelling when viewed in terms of the First and
Fourteenth Amendments and in light of the public interest in a vigorous press.”).
794 MARYLAND LAW REVIEW [Vol. 71:772

Thus, trial transcripts, most of the evidence introduced at trial,


and, in some cases, settlement agreements, 107 have been found to
qualify as part of the court record. 108 Even documents filed in con-
junction with a denied motion for summary judgment may be subject
to public access on the basis that denial of the motion “shaped the
scope and substance of the litigation.” 109 But once again, this all va-
ries by jurisdiction and by individual court.
Exclusions by omission: In some instances, documents that might
technically be considered part of the record because they were relied
upon by the judge or jury, or introduced at trial, might nevertheless
be excluded from the official court record to which the public has
access. 110 Generally speaking, exhibits that are too cumbersome to be

107. See, e.g., SEC v. Van Waeyenberghe, 990 F.2d 845, 849 (5th Cir. 1993) (holding that
a settlement agreement filed in a district court is a judicial record to which the right of
access applies); see also In re September 11 Litig., 723 F. Supp. 2d 526, 528, 531–33
(S.D.N.Y. 2010) (same); Bank of Am. Nat’l Trust & Sav. Ass’n v. Hotel Rittenhouse As-
socs., 800 F.2d 339, 345 (3d Cir. 1986) (granting in part motion to unseal settlement
agreement, reasoning that “[t]here is no persuasive reason to believe that the public or
the press will mischaracterize the settlement information and construe it as an admission
of liability” and that “[t]he Aviation Defendants’ fear of public misperception is specula-
tive and unlikely, and does not outweigh the presumption of public access to the settle-
ment information”). But see Jessup v. Luther, 277 F.3d 926, 928–30 (7th Cir. 2002) (find-
ing that settlement agreements are not judicial records and therefore not presumptively
public, but that the particular settlement agreement at issue was a public document be-
cause it involved input from a federal judge and had already been made publicly availa-
ble). The court held that the presumption of access generally does not apply to settlement
agreements, pointing out that “[p]arties who settle a legal dispute rather than pressing it
to resolution by the court often do so, in part anyway, because they do not want the terms
of the resolution to be made public.” Id. at 928. The court further noted that defendants
may be particularly reluctant to disclose the terms of a settlement for fear that doing so
might “encourage others to sue.” Id. Because settlement agreements are not judicial
records, “the issue of balancing the interest in promoting settlements by preserving secrecy
against the interest in making public materials upon which judicial decisions are based
does not arise—there is no judicial decision.” Id.
108. See, e.g., FTC v. Standard Fin. Mgmt. Corp., 830 F.2d 404, 409 (1st Cir. 1987)
(“[W]e rule that relevant documents which are submitted to, and accepted by, a court of
competent jurisdiction in the course of adjudicatory proceedings, become documents to
which the presumption of public access applies”); Bank of Am. Nat’l Trust & Sav. Ass’n, 800
F.2d at 343 (finding that the common law presumption of access applied to a settlement
agreement between the bank and a hotel developer); Martin, 746 F.2d at 968 (acknowledg-
ing a common law right of access to evidence introduced at trial, but noting that the right
“is not limited to evidence” (citation omitted)). But see Littlejohn, 851 F.2d at 682 (holding
that because exhibits would be destroyed by the clerk if not returned to the parties, they
are no longer a part of the judicial record subject to public access).
109. Republic of the Philippines v. Westinghouse Elec. Corp., 949 F.2d 653, 660 (3d Cir.
1991).
110. See Poliquin v. Garden Way, Inc., 989 F.2d 527, 532–33 (1st Cir. 1993) (noting that
post-trial access to documents introduced at trial can be restricted “only [upon] the most
compelling showing”).
2012] ONLINE COURT RECORDS 795

kept at the clerk’s office, or that are not in a form readily available for
copying, might be omitted from the record. 111 In what is perhaps an
extreme example, the Third Circuit noted in Littlejohn v. Bic Corp. 112
that “an animate object, such as a member of an endangered species,”
presents problems for the clerks, since “[i]n such circumstances, a
continuing duty to retain trial exhibits is untenable because it would
require the Clerk or the owner to maintain the endangered creature
for the remainder of its life.” 113
A list of juror names will likely be omitted from the record,
though this does necessarily mean that a curious member of the press
or public could not find out who served. 114 In New Jersey, while a list
of juror names is not disclosed, 115 the voir dire (jury selection) tran-
script is publicly available. 116 In most cases, the jurors will be called by
their full names during voir dire, prior to receiving juror numbers.117
Unless the transcript is sealed, juror names are publicly available in
New Jersey, although these names may be difficult to ascertain if the
final panel is not clearly identified in the transcript.
No matter how indirect, the inclusion of jurors’ names in the
record is particularly disconcerting. 118 As Peter Winn has pointed out,
“Court records contain sensitive information about crime victims,
about witnesses, about jurors, and about a host of other persons[,
and] [t]he parties to a dispute may ignore the information privacy in-
terests of third parties . . . .” 119 As explained above, the burden is al-
most always on the parties or their lawyers to redact sensitive or per-

111. E.g., NEB. R. 6-501, 6-502; WASH. R. APP. P. 9.8(b).


112. 851 F.2d 673 (3d Cir. 1988).
113. Id. at 682 n.23.
114. See KEYES, supra note 25 (discussing a federal court policy that “mandates that doc-
uments containing identifying information about jurors . . . are no longer included in the
public case file,” but noting that jurors could be identified “through jury selection” or
“simply by digging through the case file”).
115. Id. But see N.J. STAT. ANN. § 2B:20-5 (West 2011) (“The list of names randomly se-
lected from the juror source list shall be filed and publicly posted . . . .”).
116. United States v. Antar, 38 F.3d 1348, 1351 (3d Cir. 1994); Barber v. Shop-Rite of
Englewood & Assocs., Inc., 923 A.2d 286, 292–93 (N.J. Super. Ct. App. Div. 2007).
117. See N.J. STAT. ANN. § 2B:23-10 (West 2011) (stating that parties “may question any
person summoned as a juror after the name is drawn and before the swearing”); Antar, 38
F.3d at 1351 (noting that during voir dire, “the members of the petit jury stated their names
and hometowns on the record”).
118. See Nancy J. King, Nameless Justice: The Case for the Routine Use of Anonymous Juries in
Criminal Trials, 49 VAND. L. REV. 123, 127–28 (1996) (supporting the argument in favor of
having jurors serve anonymously by noting that “enough harassment by opponents of ver-
dicts takes place to keep many jurors worried”).
119. Winn, supra note 95, at 151 (footnotes omitted).
796 MARYLAND LAW REVIEW [Vol. 71:772

sonal information. 120 For those individuals who are involved in a law-
suit but are not represented—either because they are acting pro se or
because they are only collaterally involved in the action—it may be
the case that no one remembers or bothers to redact their informa-
tion. 121
Before addressing the role of courts in making “good cause” de-
terminations and balancing the public’s right of access to court
records against individual privacy concerns, 122 it bears noting that
these carefully crafted judicial rules and state statutes, directing which
parts of the case file become part of the publicly accessible court
record, may at times, in practice, recede into the background. 123
Clerks and courthouse employees, overwhelmed with paperwork and
record requests, are likely to rely just as much on custom and conven-
ience as on rules handed down from higher courts, particularly when
only a few records are requested at a time. 124 Often, it is not practica-
ble for one or two employees to sift through hundreds or even thou-
sands of records, removing Social Security numbers, financial data, or
other information that has been left in the record by the parties or
their lawyers. 125 When these records are not already in electronic

120. FED. R. CIV. P. 5.2(a).


121. See Peter A. Winn, Online Court Records: Balancing Judicial Accountability and Privacy
in an Age of Electronic Information, 79 WASH. L. REV. 307, 312, 321 (2004) (noting that
“courts are sensitive to protect . . . the harm that can come to . . . third parties, who may
have no control over the information so disclosed[,]” and who may have “never intended”
that their information would be easily accessible in an electronic record).
122. See infra Part II.C.
123. See Sudbeck, supra note 105, at 89–90 (explaining that “[t]raditionally, access issues
have been determined by judges . . . or by the control the clerk exercises, . . . as custodian
of these records” and that without an “extensive body of case law to draw from, court ad-
ministrators, not judges, find they must develop policies for access to records that will pro-
tect private and sensitive information”).
124. See Martin, supra note 4, at 874–76 (explaining that a statewide uniform system of
case management “is a distant prospect” in many states and using California as an example
of the negative effects of this “decentralization” because of the vastly different approaches
that exist in courts throughout the state). When bulk record access is requested, this is
more likely to come to the attention of the judges at the courthouse. See, e.g., Burnett v.
County of Bergen, 968 A.2d 1151, 1164 (N.J. 2009) (holding that “bulk disclosure of realty
records to a company planning to include them in a searchable, electronic database” was
acceptable if plaintiffs paid for redaction of Social Security numbers); Higg-A-Rella, Inc. v.
County of Essex, 660 A.2d 1163, 1166 (N.J. 1995) (holding that plaintiffs have a common-
law right to obtain electronic tax-assessment data for “every parcel of real estate in each of
the county’s municipalities”). In addition, N.J. R. 1:38-7(f) states that “[a]ny request for
the mass release, in bulk, of electronically stored or microfilmed records containing Social
Security numbers must be submitted to the Administrative Director of the Courts. A fee
may be charged for the cost of redacting Social Security numbers from such records.”
125. See Winn, supra note 121, at 320–21 (arguing that “case files often contain private
or sensitive personal information” for which there will be no protection when the files be-
2012] ONLINE COURT RECORDS 797

form, the job of removing sensitive information becomes much more


burdensome. 126 When they are in electronic format, court clerks may
in some instances, without oversight, decide to simply place all the
records online to avoid having to complete paper requests at the
courthouse and to provide greater accessibility to interested parties. 127
Thus, court decisions addressing what parts of the case file are and
are not available for public consumption should be treated as only
part of the equation that will determine the actual level of access at
the moment that a request is made.

C. Balancing the Presumption of Access Against Other Considerations


While “there is a ‘strong presumption in favor of public access to
judicial proceedings,’” 128 it may be overcome based on the following
six factors:
(1) the need for public access to the documents at issue;
(2) the extent of previous public access to the documents;
(3) the fact that someone has objected to disclosure, and the
identity of that person;
(4) the strength of any property and privacy interests as-
serted;
(5) the possibility of prejudice to those opposing disclosure;
and
(6) the purposes for which the documents were introduced
during the judicial proceedings. 129

come accessible electronically); Schwartz, supra note 14 (discussing how a spokeswoman


from the Administrative Office of the U.S. Courts stated that “courts comb through the
documents on a regular basis” but that a search of federal court documents from Washing-
ton state found that “thousands of documents” had not been properly redacted (internal
quotation marks omitted)); see also N.J. R. 1:38-7(c)(1) (“[P]arties shall certify in the Case
Information Statement that all confidential personal identifiers have been redacted and
that subsequent papers submitted to the court will not contain confidential personal iden-
tifiers in accordance with the provisions of this rule.”).
126. See supra text accompanying notes 41–42.
127. See Jennifer 8. Lee, Dirty Laundry, Online for All to See, N.Y. TIMES, Sept. 5, 2002, at
G1, available at http://www.nytimes.com/2002/09/05/technology/dirty-laundry-online-
for-all-to-see.html (discussing how policies at the local levels for online court documents
“have tended to be murky or nonexistent” and can “sometimes fall[] to a single person”
and providing as an example the clerk of courts for Hamilton County, Ohio, who, with his
technology staff, created a website to provide online access to documents, such as “[s]tate
tax liens, arrest warrants, [and] bond postings,” which were already scanned in electronic
form).
128. In re Sealed Case, 237 F.3d 657, 666 (D.C. Cir. 2001) (quoting Johnson v. Greater
Se. Cmty. Hosp. Corp., 951 F.2d 1268, 1277 (D.C. Cir. 1991)).
129. EEOC v. Nat’l Children’s Ctr., Inc., 98 F.3d 1406, 1409 (D.C. Cir. 1996) (citing
United States v. Hubbard, 650 F.2d 293, 317–22 (D.C. Cir. 1980)).
798 MARYLAND LAW REVIEW [Vol. 71:772

Each of these factors can trigger a balancing inquiry in which a court


weighs the public’s interest in open access against countervailing as-
serted interests. 130
There appears to be a general agreement among courts that the
right of access to court records ensures a well-informed public, 131 pro-
vides an “educational and informational benefit” to the citizenry, 132
and allows the public to monitor the functioning of the courts, 133 the-
reby providing an additional check on the judicial system and increas-
ing judges’ accountability. 134 According to the Second Circuit, the
right of access to judicial documents “is based on the need for federal
courts, although independent—indeed, particularly because they are
independent—to have a measure of accountability and for the public
to have confidence in the administration of justice.” 135
As we have explained above, items such as sealed discovery mate-
rials 136 and financial information137 submitted by the parties are gen-
erally not included in the court record, due primarily to privacy con-
cerns and the belief that allowing access to such documents would not
serve any of the purposes of public access. 138 As an Eastern District of
Pennsylvania court stated in Zenith Radio Corp. v. Matsushita Electric In-
dustrial Co.:

130. See id. at 1408–10 (reversing the district court’s decision to seal a consent decree
after weighing the six factors and finding “that on balance the nature of the services pro-
vided by the Center as well as the Center’s receipt of public funding cuts against rather
than in favor of sealing the record”).
131. United States v. Mitchell, 551 F.2d 1252, 1258 (D.C. Cir. 1976) (“[T]he right of
inspection serves to produce ‘an informed and enlightened public opinion’” (quoting
Grosjean v. Am. Press Co., 297 U.S. 233, 247 (1936)), rev’d, Nixon v. Warner Commc’ns,
Inc., 435 U.S. 589 (1978).
132. United States v. Criden, 648 F.2d 814, 829 (3d Cir. 1981).
133. Mitchell, 551 F.2d at 1258; see also In re Cont’l Ill. Sec. Litig., 732 F.2d 1302, 1313–14
(7th Cir. 1984) (noting that “factors weighing in favor of public disclosure of court docu-
ments” include “‘the public’s interest in assuring that the courts are fairly run and judges
are honest’” (quoting Crystal Grower’s Corp. v. Dobbins, 616 F.2d 458, 461 (10th Cir.
1980))).
134. See Globe Newspaper Co. v. Superior Court, 457 U.S. 596, 606 (1982) (discussing
the check on accountability in the context of criminal trials).
135. United States v. Amodeo, 71 F.3d 1044, 1048 (2d Cir. 1995).
136. See, e.g., Leucadia, Inc. v. Applied Extrusion Techs., Inc., 998 F.2d 157, 163–65 (3d
Cir. 1993) (holding that the common-law right of access does not extend to sealed discov-
ery materials).
137. See, e.g., United States v. Lexin, 434 F. Supp. 2d 836, 849 (S.D. Cal. 2006) (finding
“that the documents containing Defendants’ personal financial information submitted to
support their Requests for Appointment of Counsel, are not judicial documents” and
therefore are not subject to public disclosure).
138. See supra Part II.B.2.c; see also Leucadia, 998 F.2d at 164 (noting that discovery was
not open to the public at common law and is conducted in private as a matter of modern
practice) (citing Seattle Times Co. v. Rhinehart, 467 U.S. 20, 33 (1984)).
2012] ONLINE COURT RECORDS 799

[W]hether various kinds of materials are part of the judicial


record is a question that must be answered for each in light
of the purposes served by the common law right to inspect
and copy public records[,] [b]earing in mind that access
rights exist to promote knowledge of and attention to the
performance of the courts, for the benefit of society as a
whole . . . . 139
The Zenith Radio court went on to suggest that the determina-
tions of whether particular information is considered part of the court
record to which the presumption of public access applies must be
made on a case-by-case basis, considering “the magnitude and immi-
nence of the threatened harm from disclosure, including the particu-
larized interests of the litigants against disclosure . . . and the availabil-
ity of less restrictive alternatives.” 140 The language courts use in
balancing these competing interests is broad and malleable; judges
often make statements, for instance, that a district court should deny
public access only if, “after weighing the interests advanced by the
parties in light of the public interest and the duty of the courts . . . jus-
tice so requires.” 141 In the context of criminal trials, the United States
Supreme Court has held that certain “features of the criminal justice
system” weigh heavily in favor of public access to criminal trials. 142 In
order to overcome this presumption of access and “inhibit the disclo-
sure of sensitive information,” the moving party must show that the
denial of access “is necessitated by a compelling governmental inter-
est, and is narrowly tailored to serve that interest.” 143
When a party moves to have all or a portion of the record sealed,
the court must decide whether to grant this motion by weighing the

139. 529 F. Supp. 866, 898 (E.D. Pa. 1981); see also Amodeo, 44 F.3d at 145 (“[T]he mere
filing of a paper or document with the court is insufficient to render that paper a judicial
document subject to the right of public access. . . . [T]he item filed must be relevant to the
performance of the judicial function and useful in the judicial process in order for it to be
designated a judicial document.”).
140. Zenith Radio Corp., 529 F. Supp. at 912 (footnote omitted).
141. In re Nat’l Broad. Co., 653 F.2d 609, 613 (D.C. Cir. 1981) (footnote and internal
quotation marks omitted).
142. Globe Newspaper Co. v. Superior Court, 457 U.S. 596, 605–06 (1982) (noting that
“the right of access to criminal trials” (1) is firmly rooted in our country’s legal history and
(2) “plays a particularly significant role in the functioning of the judicial process”).
143. Id. at 606–07. Courts do not necessarily follow this in practice when deciding
whether to seal all or part of the court record. See Ken Armstrong et al., The Cases Your
Judges Are Hiding From You, SEATTLE TIMES, Mar. 5, 2006, at A1, available at
http://seattletimes.nwsource.com/html/yourcourtstheirsecrets/2003595519_webseal05m
NEW.html (discussing a survey of civil suits filed in the King County Superior Court in
Washington state and arguing that “judges have displayed an ignorance of, or indifference
to, the legal requirements for sealing court records”).
800 MARYLAND LAW REVIEW [Vol. 71:772

party’s interest in confidentiality against the public interest in access


to court records. In both state and federal court, a party may move to
limit access to a portion of or even an entire record upon a showing
of “good cause.” 144 According to the New Jersey rules of court:
Good cause to seal a record shall exist when: (1)
[d]isclosure will likely cause a clearly defined and serious in-
jury to any person or entity; and (2) [t]he person or entitys
[sic] interest in privacy substantially outweighs the presump-
tion that all court and administrative records are open for
public inspection . . . . 145
The Federal Rules of Civil Procedure, as well as many states, do not
define “good cause” but have left it to the courts to interpret. 146
State and federal courts have taken a variety of approaches in ba-
lancing “good cause” to seal court records against the presumption of
public access, from requiring that court proceedings “be open to the
fullest public scrutiny,” unless the law provides otherwise, 147 to sug-
gesting (rather ambiguously) that the public “right of access may be
overcome when a sufficiently compelling interest for nondisclosure is
identified.” 148 Courts frequently emphasize the role of judicial discre-
tion in determining on a case-by-case basis whether a particular
record, or a portion thereof, is to be sealed, rather than identifying
whole categories of information to exclude in advance, or laying
down bright-line rules stating what is and is not to be included in the
court record. 149

144. N.J. R. 1:38-11(a) (allowing information to be sealed for good cause); FED. R. CIV.
P. 5.2(e) (permitting redaction and limitations on “a nonparty’s remote electronic access”
for good cause).
145. N.J. R. 1:38-11(b).
146. FED. R. CIV. P. 5.2(e) (requiring good cause for redactions and limitations on non-
parties’ remote access); OR. REV. STAT. § 135.873(2) (2009) (sealing records when seeking
to obtain protective orders); VT. STAT. ANN. tit. 33, § 5119(a)(3) (Supp. 2011) (sealing
records in juvenile cases); CAL. R. APP. P. 8.46(e)(2) (sealing records on appellate review).
Furthermore, the burden of persuasion is on the moving party in federal court. See Bank of
Am. Nat’l Trust & Sav. Ass’n v. Hotel Rittenhouse Assocs., 800 F.2d 339, 344 (3d Cir. 1986)
(“The burden is on the party who seeks to overcome the presumption of access to show
that the interest in secrecy outweighs the presumption.”).
147. N.Y. Post Corp. v. Leibowitz, 143 N.E.2d 256, 258 (N.Y. 1957).
148. Associated Press v. State, 888 A.2d 1236, 1245 (N.H. 2005).
149. See Nixon v. Warner Commc’ns, Inc., 435 U.S. 589, 599 (1978) (“The few cases that
have recognized [the public’s right to access court records] do agree that the decision as
to access is one best left to the sound discretion of the trial court, a discretion to be exer-
cised in light of the relevant facts and circumstances of the particular case.”). But see N.J.
R. 1:38-1 (“Court records and administrative records . . . within the custody and control of
the judiciary are open for public inspection and copying except as otherwise provided in
2012] ONLINE COURT RECORDS 801

By and large, courts are reluctant to seal records in the face of


the strong presumption of open access. Thus, when an individual or
her lawyer seeks to deny the public access to a particular record, she
must demonstrate not only good cause, but “‘an overriding interest
based on findings that closure . . . is narrowly tailored to serve that in-
terest.’” 150 While a court may agree to seal a particular document in
the record that contains a great deal of personally identifiable infor-
mation, many courts are reluctant to seal the entire record, because
doing so imposes “a burden on the courts and [is] an extreme incon-
venience to attorneys.” 151 Even when adverse parties stipulate to the
sealing of records, “the [c]ourt must weigh any interests in confiden-
tiality against that of the public to open court records.” 152 Additional-
ly, “[a] record should be sealed in its entirety only in extremely li-
mited situations.” 153 Courts frequently cite a fear of intruding on a
constitutionally protected right to access court records, extra burdens
placed on the courts, and inconvenience to attorneys as reasons to
seal records only in very few circumstances. 154 Claims that a court
deems less than compelling will generally not provide good cause to
seal otherwise open records. 155 The Colorado Court of Appeals has
suggested that because injury to one’s reputation is an inherent risk

this rule. Exceptions enumerated in this rule shall be narrowly construed in order to im-
plement the policy of open access to records of the judiciary.”).
150. Publicker Indus., Inc. v. Cohen, 733 F.2d 1059, 1071 (3d Cir. 1984) (quoting Press-
Enterprise Co. v. Superior Court, 464 U.S. 501, 510 (1984)); see also Globe Newspaper Co.
v. Superior Court, 457 U.S. 596, 606–07 (1982) (holding that even in criminal cases, if “the
State attempts to deny the right of access in order to inhibit the disclosure of information,
it must be shown that the denial is necessitated by a compelling governmental interest,
and is narrowly tailored to serve that interest”).
151. Davis v. Davis, 997 So.2d 149, 161 (La. Ct. App. 2008); see also In re Orion Pictures
Corp., 21 F.3d 24, 26 (2d Cir. 1994) (finding that “on a purely practical level, the sealing
of court records inflicts a costly nuisance on the judicial system . . . [and] impose[s] sub-
stantial burdens on the clerk’s office and on a judge’s staff” (citation omitted)).
152. Carty v. Virgin Islands, 203 F.R.D. 229, 230 (D.V.I. 2001).
153. Davis, 997 So.2d at 161.
154. Id. at 160. When the reputations of unrepresented third parties are at stake, the
balancing may—or perhaps should—come out differently. See supra Part II.C (discussing
the privacy concerns of unrepresented third parties).
155. See, e.g., Kamakana v. City of Honolulu, 447 F.3d 1172, 1183 (9th Cir. 2006) (hold-
ing that claimed reliance by a non-party on a stipulated protective order was not a compel-
ling reason sufficient to overcome the presumption of public access); Va. Dep’t of State
Police v. Washington Post, 386 F.3d 567, 580 (4th Cir. 2004) (holding that the State Police
offered no compelling reason to keep documents under seal).
802 MARYLAND LAW REVIEW [Vol. 71:772

in almost every civil suit, such concerns should not overcome the
strong presumption in favor of open access to court records. 156
Some circumstances so predictably raise sufficient privacy con-
cerns that courts will almost always grant a party’s request to seal the
record. For example, courts are likely to find good cause to seal
records when children or victims of sexual assault are involved, 157 or
when the records in question were previously sealed and used only
during discovery. 158 In other words, a judicial finding of good cause
often turns on the very same factors that the New Jersey court ac-
counts for statutorily in Rule 1:38. The reasons for this are relatively
straightforward: While encouraging people to use the justice system is
frequently cited as a reason for keeping court records open, in cases
where good cause has been found to seal records, judges often note
that sealing the record will actually make people more willing and
likely to use the system by allaying their fears about the exposure of
sensitive personal information or business trade secrets. 159
It bears noting that when courts weigh the public’s interest in
access against an individual’s interest in privacy, they tend to focus on
issues of defamation, 160 disclosure of trade secrets, 161 and, in the case

156. Doe v. Heitler, 26 P.3d 539, 544 (Colo. App. 2001) (finding “that a court file con-
tain[ing] extremely personal, private, and confidential matters is generally insufficient to
constitute a privacy interest warranting the sealing of the file”).
157. Application of Lascaris, 319 N.Y.S.2d 60, 62–63 (N.Y. Sup. Ct. 1971) (holding that
court files relating to a child neglect case were to be kept confidential to protect the par-
ties and to encourage resolution of family problems through legal means).
158. Leucadia, Inc. v. Applied Extrusion Techs., Inc., 998 F.2d 157, 163 (3d Cir. 1993)
(“‘[T]here is no tradition of public access to discovery, and requiring a trial court to scru-
tinize carefully public claims of access would be incongruous with the goals of the discov-
ery process.’” (quoting Anderson v. Cryovac, Inc., 805 F.2d 1, 13 (1st Cir. 1986))).
159. See Lascaris, 319 N.Y.S.2d at 63 (sealing the record in a child-neglect case “to en-
courage the family resolution of problems of the family through legal means,” which is
“not likely to happen unless there is assurance that confidences will be kept”). But see
White v. Worthington Indus., Inc. Long Term Disability Income Plan, 266 F.R.D. 178, 196
(S.D. Ohio 2010) (denying the plaintiff’s motion to seal records containing medical in-
formation because the court “relied on portions of [these records] in making its decision
[and] the public’s entitlement to view court documents upon which a court’s decision
rests ordinarily outweighs any privacy interest in such records”). In White, the court noted
that “[t]he plaintiff who seeks such redress ordinarily understands that in order to do so,
he or she may be waiving the right to keep his or her medical history out of the public
domain, and Ms. White’s opposing memorandum acknowledges that she appreciates this
eventuality.” Id. Thus, while recognizing that this principle undermined the plaintiff’s
interest in privacy, the court nonetheless found in favor of public access based, at least in
part, on the fact that the plaintiff chose to undertake this risk when she began the litiga-
tion. Id.
160. See In re Food Mgmt. Grp., LLC, 359 B.R. 543, 561 (Bankr. S.D.N.Y. 2007) (stating
that the court must seal court records if the interested party “can demonstrate that the al-
legations in the complaint are in fact scandalous and defamatory”); cf. Brown & William-
2012] ONLINE COURT RECORDS 803

of special victims, physical or emotional safety. 162 What is rarely, if ev-


er, included in this balancing are concerns about the kind of personal
information discussed in Part II.A. This information can be easily ag-
gregated and linked to particular individuals by companies that traffic
in the free collection and highly profitable sale of large amounts of
personal data, 163 or worse, by identity thieves. Since the scale already
tips so far in favor of open access, it would be hard to imagine a judge
finding that the risk of disclosing an unredacted Social Security num-
ber outweighs the strong presumption of public access. 164 This com-
pounds the risk posed by the vast amount of personal information
readily available to anyone with an interest in finding it—whatever
their intentions.

III. CONTEXTUAL INTEGRITY


In the preceding Part, we provided an overview of the legal right
of access to court records, as well as recognized grounds, such as pri-
vacy, for curtailing access. This landscape is where we begin our anal-
ysis of whether and how these rights and corresponding obligations

son Tobacco Corp. v. FTC, 710 F.2d 1165, 1179 (6th Cir. 1983) (finding that harm to a
company’s reputation does not by itself justify sealing the record); Nicklasch v. JLG Indus.,
Inc., 193 F.R.D. 570, 574 (S.D. Ind. 1999) (finding that possible embarrassment to a com-
pany is not enough to seal court records).
161. See Valley Broad. Co. v. U.S. Dist. Ct., 798 F.2d 1289, 1294 (9th Cir. 1986) (“Coun-
seling against [public] access would be the likelihood of an improper use, ‘including pub-
lication of scandalous, libelous, pornographic, or trade secret materials . . . .’” (quoting
United States v. Criden, 648 F.2d 814, 830 (3d Cir. 1981) (Weis, J., concurring and dissent-
ing))); In re Iowa Freedom of Info. Council, 724 F.2d 658, 664 (8th Cir. 1983) (holding
that a private party’s property interests in its trade secrets sufficed to override the public’s
right of access because “trade secrets partake of the nature of property, the value of which
is completely destroyed by disclosure”).
162. See United States v. Cojab, 996 F.2d 1404, 1408 (2d Cir. 1993) (“[W]e have recog-
nized as additional sufficient reasons for closure and sealing those occasions where . . .
publicity might put at risk the lives or safety of government agents engaged in undercover
activities.”); Gannett Co. v. Burke, 551 F.2d 916, 916 (2d Cir. 1977) (“[I]t may be proper
for a court to seal certain records or papers, the revelation of which might, for example,
endanger a witness’s safety[.]”).
163. Helen Nissenbaum, Privacy as Contextual Integrity, 79 WASH. L. REV. 119, 121 (2004)
[hereinafter Nissenbaum, Contextual Integrity] (“Personal data is the ‘gold’ of a new catego-
ry of companies, like Axciom, that sell this information, sometimes organized by individual
profiles, to a variety of parties . . . .”).
164. See, e.g., Ostergren v. Cuccinelli, 615 F.3d 263, 286–87 (4th Cir. 2010) (holding that
Virginia could not punish, under its privacy statute, an individual who obtained land
records with unredacted Social Security numbers and published the records online).
804 MARYLAND LAW REVIEW [Vol. 71:772

are affected by the placement of digitized records online, in particu-


lar, through the lens of contextual integrity. 165
The theory of contextual integrity accounts for a right to privacy
in personal information (that is, information about persons) in terms
of appropriate flow. 166 Instead of characterizing privacy as control
over personal information, or as the limitation of access to informa-
tion, it characterizes privacy as conformance with appropriate flows of
information, in turn modeled by the theoretical construct of context-
relative (or context-specific) informational norms. When information
is captured or disseminated in ways that violate informational norms,
privacy as contextual integrity is violated.
The theory is contextual because social contexts are taken as or-
ganizing principles of social life. As such, people act and interact not
simply as individuals in an undifferentiated social world but as indi-
viduals in certain capacities in a plurality of distinct social realms.
These realms, or contexts, are structured social settings characterized
by distinct configurations of roles, activities and practices, purposes
and values, and context-specific norms prescribing expected beha-
viors. 167 Living in modern industrial societies, familiar contexts in-
clude healthcare, the marketplace, finance, politics, religion, educa-
tion, friends, and home life.
Of special relevance to contextual integrity is a subset of norms—
those that govern the flows of personal information and shape our
expectations of privacy. The theory labels these context-relative in-
formational norms and claims that their typical form is characterized
by three key variables: actors (subjects, senders, and recipients), 168 in-
formation types, 169 and transmission principles. 170 Accordingly, con-

165. See generally NISSENBAUM, PRIVACY IN CONTEXT, supra note 21 (setting out the
theory and framework of contextual integrity).
166. The “flow” of information is defined as the “transmission, communication, trans-
fer, distribution, and dissemination—from one party to another, or others.” Id. at 140.
167. The idea of social life as constructed from differentiated realms is not invented in
the theory of contextual integrity. Rather, it is drawn and generalized from rigorous work
by social theorists and philosophers who have put forth similar ideas, including domains,
institutions, and fields in their respective theories.
168. Information senders and recipients form the two poles of information transmission—
one sends, the other receives. These may be individuals, groups, or entities like organiza-
tions or committees. NISSENBAUM, PRIVACY IN CONTEXT, supra note 21, at 141. Informa-
tion subjects, however, are generally individuals. Id. They are the actors to whom the in-
formation refers, quite literally, the subjects of the information transferred.
169. Information types, or attributes, refer to the nature of the information in question:
not only who it was about, and to whom and from whom it was shared, but what it was
about. Id. at 143.
2012] ONLINE COURT RECORDS 805

text-relative information norms prescribe flows of personal informa-


tion based on the type of information in question, the capacities in
which subjects, senders, and recipients of information are acting, and
the terms under which information is conveyed—also called “trans-
mission principles.” Transmission principles define constraints on in-
formation flows. 171 Although not typically explicated in other ap-
proaches to privacy, this variable is significant in dictating whether
information may be shared only voluntarily; with due notice; bought,
sold, or shared without cost; shared in confidence; entrusted to a
third party for use only in an information subject’s interest; extracted
under duress; shared only pursuant to a warrant; and so on. 172
It is important to note as a point of comparison between contex-
tual integrity and other privacy theories that all the variables matter.
This means, for example, that no information is totally private or to-
tally public but is always defined relative to the actors and transmis-
sion principles. 173 (Whenever we omit mention of any of the va-
riables, it is usually possible to spell these out; unfortunately elliptical
speech is sometimes the root of much misunderstanding over the na-
ture of privacy expectations.) When norms are respected, privacy as
contextual integrity is preserved. It is when an action or practice con-
travenes a norm in one or more of the variables that our privacy ex-
pectations are violated. Fears and complaints that privacy has been
violated frequently follow the deployment of novel socio-technical sys-
tems, particularly those involving information technology and digital
media, because these have been frequent sources of radical disrup-
tions in information flows. 174
If we conclude that particular actions or practices violate rights to
privacy based solely on whether they contravene entrenched norms,

170. Transmission principles function to constrain “the flow (distribution, dissemina-


tion, transmission) of information from party to party in a context. The . . . terms and
conditions under which such transfers ought (or ought not) to occur.” Id. at 145.
171. For a discussion of transmission principles, see id. at ch. 7.
172. Id.
173. Nissenbaum, Contextual Integrity, supra note 163, at 139 (adding that even in the
most public of places, there may be an appropriate expectation of privacy relative to in-
formational norms).
174. Information technology in particular is inherently “socio-technical” and is guided
by the social and cultural norms of society. See Lauren Gelman, Privacy, Free Speech, and
“Blurry-Edged” Social Networks, 50 B.C. L. REV. 1315, 1319 (2009) (discussing the problems
of protecting privacy in the age of social networks and the prevalence of online divulgence
of personal information); Katherine J. Strandburg, Home, Home on the Web and Other Fourth
Amendment Implications of Technosocial Change, 70 MD. L. REV. 614, 673 n.290 (2011) (de-
scribing, for example, privacy disputes arising from changes to Facebook and Google set-
tings as socio-technical changes).
806 MARYLAND LAW REVIEW [Vol. 71:772

the resulting account would be unhelpfully conservative. Although it


would provide a refined way to detect and identify change, the ac-
count would offer no way to distinguish good change from bad, or de-
sirable, legitimate change from undesirable, unjustifiable change.
Such an account would have limited explanatory and normative pow-
er. The theory of contextual integrity addresses this problem with an
evaluative layer that allows an action or practice to be judged on mor-
al and political grounds, or, in the case of disruptive technologies and
other engines of change that challenge norms, a comparison in moral
and political terms between entrenched norms and disruptive prac-
tices. 175
The evaluative layer of contextual integrity itself involves two ana-
lyses. The first calls for an inquiry into implications of a given infor-
mation flow with particular attention to those with moral or political
significance. A flow in question might cause harm, might unbalance a
desirable power equilibrium, result in unfair discrimination, suppress
liberty and autonomy, and so on. 176 Alternatively, it might do all these
things to a greater or lesser extent than a novel flow or a flow pre-
scribed by entrenched norms with which it is compared. Here, we
draw on a great body of scholarly work discussing the value of privacy
to individuals and societies. 177
A second analysis considers the comparative impacts of novel and
entrenched flows in light of context-specific ends, purposes, and val-
ues. For example, if one of the purposes of a healthcare context is to
reduce physical suffering and its value is to achieve this purpose with-
out regard for wealth, norms would require that rich and poor alike
be relieved of pain. Additionally, to reduce the chance of prejudicial
treatment, norms might even require that information about wealth
status not be shared with caregivers. The second analysis can also
serve to break ties or resolve conflicts that occur in the first layer of
analysis. Thus, while property rights of air travelers are curtailed by

175. NISSENBAUM, PRIVACY IN CONTENT, supra note 21, at 140 (“The framework of con-
textual integrity maintains that the indignation, protest, discomfit, and resistance to tech-
nology-based information systems and practices . . . invariably can be traced to breaches of
context-relative informational norms. Accordingly, contextual integrity is proposed as a
benchmark for privacy.”).
176. For example, an unrestricted or readily available flow of information could lead to
discrimination on the basis of sexual orientation, religious beliefs, and political affiliations,
to name a few.
177. See NISSENBAUM, PRIVACY IN CONTEXT, supra note 21, at ch. 4 (describing ways in
which to determine the value in privacy); see also Nissenbaum, Contextual Integrity, supra
note 163, at 150 (noting that privacy has great importance to both individuals and society,
serving common and collective, as well as individual, purposes).
2012] ONLINE COURT RECORDS 807

luggage searches to protect the security of other travelers, also at stake


is the smooth functioning of air travel overall. All three factors,
among others, are relevant to the evaluation of norms that, in this
case, are embodied in explicit policy and regulation.
Applying the framework of contextual integrity to online place-
ment of court records means compiling an account of changes in in-
formation flows and evaluating them in moral and political terms.
Few would claim that online access to court records would have no
impact on information flows—indeed, it is precisely the prospect of
such changes that accounts for much of the enthusiastic support for
online access. 178 Some of the support seems bluntly to deny any such
change in flow, asserting that since “public is public,” a transforma-
tion from local access to online access is merely doing the same thing
more efficiently. 179 But, the “thing” in question that stays the same is
not the way information flows, as such; it is the normative commit-
ment to transparency of government functioning through open
access to court records. 180
Detractors, too, recognize that online placement creates changes
in flow. Although, to the best of our knowledge, no one has argued
flatly against the integration of digital media into court records sys-
tems, many, in writing and in numerous public discussions including
those convened by state courts, have expressed concerns over prob-
lems arising from increased exposure due to the transition from local-
ly accessible records (some still in hardcopy form) to digital records
accessible online. 181 In a New Jersey report, for example, this in-
creased exposure was labeled “hyper-dissemination,” implying that
while dissemination is clearly a good thing, because court records of-
ten contain sensitive information, hyper-dissemination requires closer
examination. 182 Similarly, others have cited the loss of practical ob-

178. See Martin, supra note 4, at 860–61 (detailing several developments that led to
widespread support for online access to court records). One could imagine other reasons,
for example, cost savings and the environment.
179. See Nissenbaum, Contextual Intergity, supra note 163, at 120–21 (asserting that plac-
ing previously available records online “is merely an administrative move towards greater
efficiency. Nothing has changed, fundamentally”).
180. See Martin, supra note 4, at 861 (acknowledging “the courts’ historic commitment
to transparency”).
181. See id. at 882–84 (describing the public’s concerns about privacy and security of
court records “especially during a period of transition”); see also Michael Caughey, Com-
ment, Keeping Attorneys from Trashing Identities: Malpractice as Backstop Protection for Clients
Under the United States Judicial Conference’s Policy on Electronic Court Records, 79 WASH. L. REV.
407, 407 (2004) (arguing that clients should be allowed to bring a malpractice claim
against their attorney who fails to redact sensitive personal information).
182. REPORT ON PUBLIC ACCESS, supra note 15, at 14.
808 MARYLAND LAW REVIEW [Vol. 71:772

scurity as a cause for caution, pointing out that the obscurity of local
records in practice has protected against unjustifiable privacy viola-
tions; they have suggested that policy and regulation may need to ad-
dress this inadvertent loss. 183
In a similar vein, our work seeks to understand and evaluate the
changes brought about by online access. Where others have invoked
concepts of hyper-dissemination and loss of practical obscurity to sig-
nal potential sources of trouble spots, the framework of contextual in-
tegrity offers a richer, more rigorous way of characterizing and eva-
luating the changes. In Part IV, we use context-relative informational
norms to characterize the nature of change brought about by digitiz-
ing records and making them available online for public access. This
involves modeling differences in flow in terms of the key parameters:
actors and information type. In Part V, we evaluate the significance of
these differences.

IV. COMPARING PATTERNS OF INFORMATION FLOW


In this Part, we systematically examine how digitizing court
records and placing them online affects flows of personal information
contained in these records. Following the framework of contextual
integrity, we model flows in terms of two key parameters: actors and
information type. 184 The relevant actors in this context are the reci-
pient, who is searching for information in court records; the subject,
whose personal information flows to the recipient; and the sender, the
information retrieval system that stores and provides access to the
court records. 185 We associate costs with flows of personal information,
where cost is an abstraction of the time, money, and effort that the
recipient has to expend to cause a flow of information. We character-
ize the effect of digitizing court records and placing them online in
terms of the difference in cost of flows of personal information in the
resulting systems as compared to paper and digital records available at
courthouses. In prior work on contextual integrity, differences in
flows caused by the introduction of new technology were characte-
rized using a binary model of flows—in effect, identifying new flows

183. Id. at 34 (“While such personal identifiers in the past remained in the ‘practical
obscurity’ of the clerk’s office, with the advent of the Internet, records that are placed on-
line are now available in an instant in one’s office or home, anywhere in the world.” (foot-
note omitted)); Martin, supra note 4, at 863 (describing the charge to governing bodies to
ensure that privacy and security are maintained for online records).
184. For a description of transmission principles, see NISSENBAUM, PRIVACY IN
CONTEXT, supra note 21, at ch. 7.
185. Id. at 141–42.
2012] ONLINE COURT RECORDS 809

that arose and old flows that were no longer possible as a conse-
quence of the introduction of the technology. 186 In contrast, here we
develop the more general notion of cost differences in flows as a basis
for characterizing the effect of the introduction of new technology.
The cost model subsumes the binary model of flows: a flow that is not
possible can be modeled by assigning to it an infinite cost, and a flow
that is possible can be modeled by assigning to it zero cost. 187 We
need this additional generality in modeling flows for our study of in-
person court record access at the courthouse and online access to
court records.
We begin in Part IV.A by presenting a model that captures key
elements of information retrieval from court records. Each step in
this model has associated costs that vary based on the characteristics
of the information retrieval system (including the medium) and the
behavior of the user (recipient) who is conducting the search. We ar-
rive at this model by augmenting previously developed information-
retrieval models with additional features that were needed to model
actual searches of court records that we conducted.
Part IV.B provides an overview of our empirical study. Specifical-
ly, we searched court records using two systems that provide online
access to court records: PACER and Google Scholar. 188 In addition,
we conducted searches at two physical courthouses: the Superior
Court County Clerk’s office in Trenton, New Jersey, and the Superior
Court of New Jersey, Hudson Vicinage. Each search followed the
steps of our information-retrieval model, 189 thus providing evidence
that the model is a suitable abstraction of the process of retrieving in-
formation from court records. The example searches capture realistic
scenarios in which the user (recipient) is acting in a certain capacity
and seeks to acquire a specific type of information about the subject.
For example, in one scenario, the recipient is a potential employer of
the subject and may be interested in knowing whether the subject was

186. See id. at 143–45 (explaining how some theorists have split information types into
dichotomies).
187. Id.
188. GOOGLE SCHOLAR, http://scholar.google.com (last visited Mar. 7, 2012). Google
Scholar provides a way to perform a broad search on all the legal opinions present in its
database. “Currently, Google Scholar allows [users] to search and read published opi-
nions of US state appellate and supreme court cases since 1950, US federal district, appel-
late, tax and bankruptcy courts since 1923 and US Supreme Court cases since 1791.”
GOOGLE SCHOLAR, SCHOLAR HELP, http://scholar.google.com/intl/en/scholar/
help.html (last visited Mar. 7, 2012). In addition, it allows users “to find influential cases
(usually older or international) which are not yet online or publicly available.” Id.
189. For a discussion of the information-retrieval model, see infra Part IV.A.
810 MARYLAND LAW REVIEW [Vol. 71:772

involved in criminal cases in the last ten years and, if so, learning
about the exact nature of the crime. 190 In two other scenarios, we
consider a lawyer who seeks detailed information about a case, 191 and
a data aggregator who is interested in building dossiers containing
personal information about individuals. 192 Our findings support the
hypothesis that the costs of retrieving various types of personal infor-
mation differ significantly in the online and local access systems, pro-
vided that the user possesses certain types of background knowledge
193
about the data subject.
Finally, in Part IV.C, we use the model and the lessons learned
from the empirical study to identify the root causes of the difference
in cost of flows of personal information between online court records
and records available at courthouses.

A. Information Retrieval Model


The goal of the eight-step model described below is to capture
key elements of information retrieval from court records. The asso-
ciated costs of each step vary according to characteristics of the in-
formation retrieval system and the behavior of users conducting the
search. Later in this Part, we discuss in detail the differences in
cost. 194
1. Information need: Imagine a user who has an information-need
about a data subject. The user seeks this information in a certain con-
text in which she is acting in a certain capacity or role relative to the
data subject. As a running example, consider a user who is interested
in knowing whether a potential employee has been involved in crimi-
nal cases and, if so, learning about the exact nature of the crime.

190. See infra Part IV.B.1.a.


191. See infra Part IV.B.1.b.
192. See infra Part IV.B.1.b.
193. Caveat: We used PACER and Google Scholar in our empirical study of online sys-
tems to search for federal court records, but the physical courthouses in which we con-
ducted our searches contained only court records for the state of New Jersey. Thus, we
cannot directly compare the cost of the same search online and at the physical courthouse.
This limitation is a consequence of the fact that online systems to access court records for
the state of New Jersey, as well as many other states, are still in the process of being de-
signed and are not yet available for public use. See Martin, supra note 4, at 872 (stating that
“state court systems have been far slower and less coordinated in making th[e] transition”
from paper records to electronic filing). Thus, “state courts seriously lag behind the fed-
eral courts in this area.” Id. The searches that we conducted online and at the physical
courthouses, respectively, are, nevertheless, still comparable in many respects, for example
in how the background knowledge of users affects their findings. We hypothesize, there-
fore, that the comparison of costs remains meaningful despite this limitation.
194. See infra Part IV.C.
2012] ONLINE COURT RECORDS 811

2. Selecting information systems: The user chooses a set of informa-


tion systems to conduct her search based on her information need.
For example, she may decide to use PACER or Google Scholar or to
use a catalog at a physical courthouse or some combination of these
systems to conduct her search. The system might require the user to
provide information about the context, that is, the purpose of her
search and her relation to the data subject.
3. Getting to the information system: The user gets to the location of
the information system (online or at the physical location).
4. Querying the system: The user then formulates and issues a query
to the system based on her knowledge and the interface presented to
her. For example, she may search by name if she knows the name of
the potential employee and the interface supports searches by the
names of parties involved in court cases (as indeed PACER’s interface
does). In addition to restrictions on the nature of queries imposed by
the interface (possibly based on the role of the user), social factors
may impose additional constraints. A curious user may be more will-
ing to search online through divorce records for details about a
neighbor or acquaintance than to get this information from a court-
house, where she has to confront a clerk. 195
5. System’s response to query: The system processes the query over
the indexed database of court records and returns a set of results.
The model of a system, including the content and format of its data-
base and data, determines how the system responds to various types of
queries. For example, a system might index only a few types of infor-
mation, such as party name and year, or it might index each term in
the full text of the documents in the court record, allowing a user to
search by any word present in the documents. In presenting certain
types of information, the system may also take the relation of the user
to the data subject into account. Confidential juvenile records, for in-
stance, may be released only to the child or a named parent, and pos-
sibly to a caseworker if she can show cause (as determined by a
judge).
6. Evaluation of results: The user evaluates the results to determine
whether her information need has been met. Several factors affect
the cost of this step, including the size of the set of results and the
manner in which it is displayed. If the information need is indeed
met, the search is over and the user moves on to Step 8. The user’s

195. See Winn, supra note 95, at 152, 155, 157 (describing how the “practical obscurity”
of a paper-based system of judicial records can present more obstacles and therefore dis-
courage people without substantial interest from accessing information).
812 MARYLAND LAW REVIEW [Vol. 71:772

information need may not be met, for example, because the set of re-
sults for a search by a common name is too large and the user cannot
figure out which records are about the person of interest to her. Al-
though the user’s information need is not met at this point, her know-
ledge may have increased, thus enabling her to formulate new que-
ries. Further, at this point, the user’s information need could change
based on what she has learned thus far. In our running example, the
user might discover that the potential employee was involved in a civil
case and decide that information in that record might be relevant for
her hiring decision.
7. Query reformulation: At this point, the user might reformulate her
query based on her knowledge and choose a system to issue the new
query. The user’s decision may depend on the search interface of-
fered by the system. For example, if the user wants to search by party
name and city, she may want to use Google Scholar instead of PACER,
since PACER does not allow queries of this form. 196 The control flow
now goes back to Step 3 above and the search proceeds iteratively.
The user can choose to terminate the search at any point, either be-
cause it seems that the information need will not be satisfied or be-
cause the cost of retrieving the desired information is too high.
8. Subsequent use and dissemination: The user uses the information
acquired for a purpose and possibly disseminates it further. In our
example, the user decides whether to offer the job to the applicant
based on the information retrieved from court records. In other situ-
ations, the information acquired from court records could be disse-
minated further on the Web, via social networks, or through data ag-
gregation services, leading to the creation of other sources that a
future user could turn to in Step 2 while searching for personal in-
formation contained in court records. For example, data aggregators
might build up new sources of information by accessing court records
and sell this information at a cost that is much lower than a user
might incur if she were to try to retrieve the information herself from
court records. As another example, if court records were to be made
available online in a form accessible to indexing by Web search en-
gines (such as Google or Bing), future users could simply use search
engines to find personal information in records at a very low cost. 197

196. See PUBLIC ACCESS TO COURT ELECTRONIC RECORDS, http://www.pacer.gov/ (last


visited Mar. 8, 2012).
197. In a recent French case, a man successfully sued Google after terms related to his
past criminal record arose in a name search based on Google’s Suggest feature. See Seth
Weintraub, French Court Convicts Google CEO Eric Schmidt of Defamation, CNNMONEY (Sept.
26, 2010, 11:54 AM), http://tech.fortune.cnn.com/2010/09/26/french-court-convicts-
2012] ONLINE COURT RECORDS 813

The model of information retrieval we describe above builds on


prior work in the information retrieval and library-science literature,
including approaches proposed by Bates; 198 Marchionini; 199 Belkin,
Oddy, and Brooks; 200 and Hearst. 201 Although all of these models
consider an iterative process similar to the one we have devised, our
model considers three additional factors that are essential for the
court-records setting and contribute to our empirical investigations
and final analysis. First, our model allows that an information retriev-
al system might provide different views of the court records to users in
different roles, for example, because of confidentiality requirements
for certain types of information (such as the juvenile records dis-
cussed in Step 5 of our model). Indeed, unlike systems studied in the
information retrieval literature whose goal is to make all types of in-
formation easily accessible, systems for retrieving information from
court records have to be designed to make certain types of informa-
tion easily accessible while ensuring that other types of information
are not. Second, in Steps 2 and 7 of our model, we allow for the pos-
sibility that the user may acquire information from multiple informa-
tion systems during the search process. While information retrieval
systems are typically studied independently, court records are, in fact,
accessible via multiple systems (PACER, Google Scholar, and emerg-
ing systems from various states). An analysis of these systems in isola-
tion is woefully inadequate since a user can easily access all the sys-
tems simultaneously in the search process. Finally, by considering the

google-ceo-eric-schmidt-of-defamation. Google has stated in its policies that individuals


concerned about inappropriate information in government records being made publicly
available should contact the respective government agency or policymakers, and that it is
the government’s responsibility to ensure that records “are free of information that in-
fringes an individual’s right to privacy.” GOOGLE, Webmaster Tools Help, About Privacy,
http://support.google.com/webmasters/bin/answer.py?hl=en&answer=82301 (last visited
Jan. 10, 2012). Google has further stated, “If information is made publicly available, we
want to help our users discover and access it.” Id. Consequently, if court records were to
be made available online in an easily indexable format, the move would likely have a
strong impact on dissemination of information and create new information flows.
198. Marcia J. Bates, The Design of Browsing and Berrypicking Techniques for the Online Search
Interface, 13 ONLINE REV. 407 (1989).
199. GARY MARCHIONINI, INFORMATION SEEKING IN ELECTRONIC ENVIRONMENTS (1995).
200. N.J. Belkin, R.N. Oddy & H.M. Brooks, ASK for Information Retrieval. Part I. Back-
ground and Theory, in READINGS IN INFORMATION RETRIEVAL 299, 299–304 (Karen Sparck
Jones & Peter Willett eds., 1997).
201. Marti A. Hearst, User Interfaces and Visualization, in MODERN INFORMATION
RETRIEVAL 257–323 (Ricardo Baeza-Yates & Berthier Ribeiro-Neto eds., 1999); see also
CHRISTOPHER D. MANNING ET AL., INTRODUCTION TO INFORMATION RETRIEVAL xiv (2008)
(giving “an up-to-date treatment of all aspects of the design and implementation of systems
for gathering, indexing, and searching documents and of methods for evaluating systems,
along with an introduction to the use of machine learning methods on text collections”).
814 MARYLAND LAW REVIEW [Vol. 71:772

potential to create new sources of information from the onward dis-


semination of acquired information, our model differs from many
other information retrieval models that typically do not consider what
happens to information once a user retrieves it. Yet this step is critical
for an analysis of the cost of information flows from court records as
discussed above in Step 8.

B. Overview of Empirical Study


To understand first-hand the cost of retrieving information from
court records, we conducted an empirical study. Specifically, we
searched court records using two systems that provide online access to
court records—PACER and Google Scholar. In addition, we searched
records at two physical courthouses—the Superior Court Clerk’s of-
fice in Trenton, New Jersey, and the Superior Court of New Jersey,
Hudson Vicinage, a trial court. We describe below these systems and
the methodology we followed in conducting searches, noting that the
steps in all these searches followed our eight-part information retriev-
al model. Our main finding is that between online and local access,
there are significant differences in the cost of retrieving various types
of personal information about a data subject. We note that in order to
successfully retrieve personal information in either medium the user
has to possess background knowledge relevant to the search. We ela-
borate on exactly what types of background knowledge are useful for
low cost retrieval of personal information using court record systems
and describe scenarios in which it is reasonable to expect that users
will possess them.

1. Search Using Online Systems

a. Systems Used
We conducted a search of nineteen federal appellate court cases
using PACER. In other words, we were searching for cases that had
been appealed from a federal district court. 202 PACER, an electronic
public access service provided by the United States judiciary, contains
“case and docket information from federal appellate, district and
bankruptcy courts and the PACER case locator via the Internet.” The
PACER case-locator interface supports search for records by various

202. The record in appellate cases contains only that on which the trial court (judge or
jury) based its decision. FED. R. APP. P. 10(a). This may or may not contain everything in
the court record that is publicly available at the trial court. FED. R. APP. P. 10(c)–(d).
2012] ONLINE COURT RECORDS 815

fields, namely region, nature of suit 203 (in case of civil records), date
when the case was filed, case title, party name, and case docket num-
ber. The advanced search interface additionally allows search by case
title or range of dates for when the case was filed or closed. Once an
initial query has been issued and the results are presented, a user can
filter results by court type (for example, appellate or bankruptcy),
court name (for example, Southern District of New York or Eastern
District of Pennsylvania), year in which the case was filed, nature of
suit (for example, contract, antitrust, labor, civil rights), or case title.
A user can also browse through records based on the nature of the
suit or look at all cases that have been filed or decided within a cer-
tain time period. In order to access the search interface, a user has to
register with PACER and pay for searches. In addition to a name and
address, a user must provide a credit card number at the time of regis-
tration or billing.
In conducting these searches, we used Google Scholar as an addi-
tional resource. Google Scholar allows users to perform a broad
search across all the legal opinions archived in its database. “Current-
ly, Google Scholar allows you to search and read published opinions
of US state appellate and Supreme Court cases since 1950, US federal
district, appellate, tax and bankruptcy courts since 1923 and US Su-
preme Court cases since 1791.” 204 A user can retrieve information
through a simple keyword search. There is no fee for this service.
Users can search for opinions from state or federal courts. Within
these opinions, users can search the full text of the document or filter
the cases by year. It is also possible to see citation history. While
Google Scholar is useful for locating judicial opinions and searching
within them, unlike PACER it does not provide detailed information
about each case and does not include a docket.

b. Observations from Search Experience


Set up for searches: In conducting our searches, we needed a
“ground truth oracle” for deciding whether our search was successful.
In other words, if we were to pick party names at random (for exam-
ple, John Smith) and searched for information about them in court
records, we would have no way of knowing for sure whether we found
the correct record (that is, whether the record we found was about

203. For a list of the available “nature of suit” search categories, see
http://www.pacer.gov/documents/natsuit.pdf.
204. See GOOGLE SCHOLAR, supra note 188.
816 MARYLAND LAW REVIEW [Vol. 71:772

the John Smith we searched for or some other person with the same
name).
We used court opinions in Google Scholar as the ground truth
oracle for all nineteen cases. Specifically, we retrieved fifteen court
opinions directly from Google Scholar (chosen at random). We also
retrieved two opinions each directly from Justia 205 and Westlaw and
subsequently found those opinions also on Google Scholar. We used
different types of information available in these court opinions as our
background knowledge in conducting the searches using PACER (we
elaborate on the exact types of background knowledge used in the
next subsection). To verify whether a search on PACER was success-
ful, we compared the party names, the case title, case number, and
the court name in the final retrieved record on PACER with the in-
formation in the original court opinion in Google Scholar.
Search results: We conducted searches assuming that the user had
different types of background knowledge about the data subject. Av-
erage generalized costs for these searches were very low: approximate-
ly $3 on PACER and under ten minutes each. We found court docu-
ments with various types of personal information, including names,
complete addresses of parties, photographs, individuals’ signatures,
personal allegations, and medical information. Operationally, the
goal of each search had been to retrieve the complaint form asso-
ciated with each case.
In one set of eleven searches, we assumed that the user knew the
complete name of the data subject. This level of knowledge is rea-
sonable in many realistic scenarios: A potential employer, curious
neighbor, and a housing application processor are but a few exam-
ples. In four of these eleven searches, the results set from the search
contained exactly one court record and included an opinion that
matched the Google Scholar opinion from which we had drawn the
name, thus indicating that we found the correct record.
Upon revisiting the seven unsuccessful searches, we expanded
the user’s background knowledge to include not only the complete
name of the data subject but also the type of case and the court where
the complaint was first filed. It is reasonable to assume this level of
knowledge in several scenarios, such as a potential employer (or an
apartment owner) conducting a background check on an applicant.
Since, for example, civil and appellate courts differ by geographic re-
gion, an employer can make a reasonable guess about the court
where such cases involving the applicant were filed, if she knows the

205. JUSTIA, http://www.justia.com/ (last visited Mar. 8, 2012).


2012] ONLINE COURT RECORDS 817

cities that the applicant has resided in; 206 information about where
the applicant has lived could be inferred from her job or housing ap-
plication. The court type to use in the search can be decided based
on the type of case a user is interested in. For example, if an employ-
er is checking whether an applicant has a criminal record, then the
court type would be criminal, and if an employer or an apartment
owner wants to check an applicant’s financial background, then she
would look at bankruptcy cases. In four out of these seven searches,
the result set from the search contained only one record.
Finally, we revisited the three searches that were unsuccessful in
the previous rounds and this time assumed that the user knew the
complete name of the data subject, the court type, the court where
the case was filed, and the range of years when the case could have
been filed. The additional background knowledge about the range of
years could be determined by the date of birth or the user may only
be interested in the subject’s criminal or bankruptcy records in the
recent past.
However, the usefulness of this information depends on whether
other records exist that correspond to the range of years in which a
user is interested. For instance in one search, we were not able to
narrow our results to find the desired record on PACER. For the oth-
er two searches, the result set was reduced to less than twenty-five
records. To narrow the search results to exactly one record, we would
have needed additional information about the data subject, such as
the exact year that the case was filed. 207
In another set of nineteen searches, we assumed that the user
knew the case title and the case number corresponding to a record.
In one case, we assumed knowledge of the case title alone. The case
title contains the names of the parties involved208 along with the name
of the court and the decision year. This level of knowledge is reason-
able in several realistic scenarios, including a lawyer who is searching

206. An exception to this line of reasoning may occur when, for example, the potential
employee was involved in out-of-state litigation that did not involve any state where she has
resided.
207. This additional information could also be in the form of Social Security numbers
(“SSN”) for bankruptcy cases. In the case of an owner checking a tenant’s financial credi-
bility or an employer looking up an employee’s record, it is reasonable to assume that the
user would know the SSN for the data subject being searched.
208. If the parties are individuals, the case usually contains all parties’ first and last
names. See supra Part II (describing exceptions to this and providing information on ano-
nymous litigation). If one or more parties is a business or a government entity, the case
title will list the name of the business or entity, and may also list the name of one or more
associated individuals. FED. R. CIV. P. 17.
818 MARYLAND LAW REVIEW [Vol. 71:772

for details of a specific case, or a data aggregator who is browsing


through court opinions on Google Scholar and using information re-
trieved from there to search for additional details about the parties on
PACER. In eleven out of these nineteen cases, the last name in the
case title was not sufficient to narrow down the results to one record.
We then used the case title to search Google Scholar and learn the
complete names of all the parties. The complete names helped us
narrow our results to one record in eighteen out of our nineteen cas-
es.
Our searches demonstrate that it is feasible to recover informa-
tion at low cost if the user possesses a reasonable level of background
knowledge.

2. Search Using Physical Systems

a. Systems Used
We conducted searches at two physical courthouses—the Supe-
rior Court Clerk’s office in Trenton, New Jersey, and the Hudson Vi-
cinage, New Jersey, trial court.
The Superior Court Clerk’s office houses various documents in-
cluding judgments, foreclosures, and liens for all trial courts in the
state. To obtain copies of records, a user has to interact with a clerk
and provide identification as well as a reason for the visit. 209 By using
a computer terminal provided at the clerk’s office, users can look for
different types of documents by accessing several programs, such as
Automated Case Management System (“ACMS”) 210 to search for
judgments by party name, docket number, venue, or judgment
amount; Judiciary Electronic Filing System (“JEFIS”), 211 to search for
foreclosures by docket number only; and “Appellate Division
Search,” 212 to search for filings and decisions by party name, case
number, organization name, phone number, zip code, motion num-

209. Location of Court Records, NEW JERSEY COURTS, http://www.judiciary.state.nj.us/su-


perior/copies_court_rec.htm (last visited Mar. 8, 2012).
210. ACMS-AUTOMATED CASE MANAGEMENT SYSTEM, http://www.judiciary.state.nj.us/
isd/acms.htm (last visited Mar. 8, 2012).
211. JUDICIARY ELECTRONIC FILING SYSTEM, http://www.judiciary.state.nj.us/jefis
/index.htm (last visited Mar. 8, 2012).
212. New Jersey Courts Search Page, RUTGERS SCHOOL OF LAW, CAMDEN,
http://lawlibrary.rutgers.edu/new-jersey-courts-search-page (last visited Mar. 8, 2012).
The New Jersey Courts website only keeps the cases for ten business days and then links its
users to the Rutgers School of Law site, which serves as a source for appellate cases dating
back to 1995.
2012] ONLINE COURT RECORDS 819

ber, type of case (criminal, civil, family, agency), and category (for ex-
ample, accusation, indictment, and grand jury for criminal cases).
At the Hudson Vicinage trial court there are three different divi-
sions: civil, family, and the Children in Court (“CIC”) Unit. In the
civil division, a user can request records only by name or docket
number, not by date of birth or address. The clerks control the
process. There are copies of documents like the complaint from a
landlord-tenant dispute, a summons, and some affidavits signed by
the attorneys. If a user wants to request a record, she must do so
through the window where the clerk sits, either after searching the
terminal for the docket number, or simply by knowing the party
names or docket number when she arrives. The clerk then retrieves
the paper record and makes a copy. The user pays for the copy at the
“fee station” where the rate is seventy-five cents per page, up to a cer-
tain number of pages, with a slightly reduced per-page fee after
that. 213
In the family division, similar to the civil division, a user has to
know either a party name or docket number to perform a search. Be-
cause of the confidential nature of records available in this division,
the clerk determines whether the user may have access to a requested
record before it is made available. 214 If the record is confidential with
regard to the public but available to particular individuals (for exam-
ple, the parties), the user must show identification. But if it is a pub-
lic record (custody dispute, divorce case), no identification is needed.
The clerk retrieves a paper copy of the record and makes a copy for
the user. This division only contains records from Hudson County,
and the cost per-page is the same as in the civil division.
The CIC Unit handles presumptively confidential juvenile
records and Division of Youth & Family Services (“DYFS”) adoptions.
These records are released only to the child or named parent, and
possibly to a caseworker if she can show cause. A judge makes the “for

213. These are the rates provided to the authors when they visited the courthouse.
However, the Superior Court website provides a different rate schedule of five cents per
page for paper copies, with certified copies costing $5 for the first five pages and seventy-
five cents per page thereafter. See Copy and Authentication Fees, NEW JERSEY COURTS,
http://www.judiciary.state.nj.us/superior/copies_court_rec.htm (last visited Mar. 8, 2012).
214. See, e.g., N.J. R. 5:17-4(b) (“Social, medical, psychological, legal, and other records
of the court or family intake services, and records of law enforcement agencies, found to
be part of a juvenile-family crisis matter, shall be strictly safeguarded from public inspec-
tion and shall be made available only pursuant to N.J.S.A. 2A:4A-60 to -62. Any application
for such records shall be made by motion to the court.”).
820 MARYLAND LAW REVIEW [Vol. 71:772

cause” determination. 215 However, a release of authorization may be


signed by the child or parent to release the records to a third party.
Unlike in the family and civil divisions, there is no charge for these
records.

b. Observations from Search Experience


In order to get to the search interface, we had to travel to the
physical location of the courthouse and pass through courthouse se-
curity. 216 We conducted searches assuming that the user knew the
party name. This level of background knowledge is reasonable in a
scenario where a data aggregator is looking for information about an
individual and is aware only of the individual’s name. The entire re-
trieval process took about an hour and a half and cost about $10.
At the Superior Court Clerk’s office in Trenton, 217 we performed
a search using the party name in ACMS and were able to get one
record in the result set. We were only able to retrieve the party
names, attorney names, docket number, and judgment amount from
the record.
We also used the JEFIS program to access foreclosures. Because
JEFIS only allowed searches by docket number, date, or filing or
transaction ID, it limited the background knowledge that would be
useful for our searches. The JEFIS search ultimately produced docu-
ments related to some cases, but these documents only listed the
plaintiff’s address (and, very rarely, the defendant’s as well), the date
of foreclosure, and the various amounts of money owed. The docu-
ments did not provide any further information.
We also performed a party-name search using the “Appellate Di-
vision Search” feature. No documents were available, however, apart
from docket entries. Addresses and occasionally phone numbers of
the parties were available, though in most cases only the attorneys’
addresses were listed.
The process of getting to the interface at the trial court in Hud-
son Vicinage, similar to the court in Trenton, was time-consuming.
Further, as mentioned earlier, it was necessary to interact with clerks
at every step. Although we were able to retrieve a record in the family

215. See N.J. STAT. ANN. § 2A.82-46 (West 2011) (stating that all information in child
assault or abuse cases is “confidential and unavailable to the public” unless judge finds
good cause at a good cause hearing and after notification to victim or victim’s family).
216. The Hudson Vicinage William Brennan Courthouse is located at 583 Newark Ave-
nue, Jersey City, NJ 07306.
217. The office of the Superior Court Clerk is at 25 W. Market Street, Trenton, NJ
08625.
2012] ONLINE COURT RECORDS 821

division, it was only two pages long and appeared to have very little in-
formation. We could not access the records in the CIC Unit because
we failed to show cause. When we tried to retrieve criminal court
records at the same location, we were directed to another building re-
ferred to simply as “the Justice Center,” about a mile away. After pass-
ing a security check at the Justice Center and stating the reason for
our visit, we were redirected to yet another building. At this location
we were asked to go back to the Justice Center, where we were told
that the criminal records were at the courthouse where we had looked
at civil, family, and CIC unit records. We could not locate the crimi-
nal records room to make a request. Practical obscurity prevailed at
this courthouse.

C. Root Causes of Cost Differences in Online and Local Access to Court


Records
In this section, we identify six root causes for the difference in
cost of flows of personal information between online and local access
systems for court records. We draw on our general understanding of
searching court records as captured in the eight-part information-
retrieval model 218 and observations from the empirical study to arrive
at these results.
1. Getting to the location of the information system. The cost differ-
ence, while unsurprising, was significantly greater than we had antic-
ipated. A user can search online if she has access to a terminal with
an Internet connection and knowledge of the URL or name of the
online system (PACER, Google Scholar, etc.). In contrast, the cost of
a physical search is greater because of the time required to get to a
courthouse and the expense of the commute. Moreover, while
records of different types are easily identifiable and searchable online,
our search for different types of cases at physical courthouses involved
hunting for the right room or building, thus adding to the cost.
2. Query interface and indexing mechanism. The query interface de-
termines what types of background knowledge can be used in the
search process. 219 Online systems, such as Google Scholar, which
support keyword search and full-text indexing, significantly reduce
the cost of information retrieval for users who may not have the exact
knowledge required to formulate queries on systems, such as PACER,

218. See supra Part IV.A.


219. See supra Part IV.A.4.
822 MARYLAND LAW REVIEW [Vol. 71:772

which allow only fixed format queries. 220 In the physical courthouses
we visited, the query interfaces were much more restrictive; 221 it was
difficult to access records for which we did not know exactly what we
were looking.
3. Linking multiple information sources. Multiple online informa-
tion sources can be queried during a search. In our experiments with
PACER, for example, we used Google Scholar as an auxiliary source
to drastically cut down on the search costs. In one test search, where
we assumed knowledge of the case title alone, a search by only the
party’s last name (as obtained from the case title) yielded more than
200 records in PACER. Searching on the case title on Google Scholar
also yielded multiple results, but when including the case year given
in the case title, we were able to pinpoint the correct record. From
this record, we obtained the case number as well as the complete par-
ty names. Returning to query PACER with the case number we were
able to find the correct court record. These searches also illustrate
that restrictions in the query interface of one search system may be
circumvented by the use of another search system. In contrast, shut-
tling between different physical courthouses to execute such linking
incurs a significantly greater cost.
4. Access restrictions. For certain types of court records, the rela-
tion between the user and the data subject determines whether access
is permitted. As discussed previously, the CIC Unit releases juvenile
records only to the child or named parent, and possibly to a case-
worker if she can show cause. It is unclear how to enforce such access
restrictions online without an expressive identification and authenti-
cation infrastructure—that is, a system that enables users to demon-
strate who they are and how they are related to the data subject. It is
technically feasible to construct such a system (for example, by using
so-called “trust management systems” 222), although the cost of deploy-
ing and maintaining such a system could be quite high. Currently,
Google Scholar does not require any form of identification and is in-
capable of providing access to such records. While PACER does re-
quire identification and authentication, it is not clear whether this in-

220. Depending on the type of case, PACER allows users to search “by case number,
party name, complete or last four digits of a social security number, case filing dates, and
much more.” PACER USER MANUAL FOR ECF COURTS, at 18 (June 2010), available at
http://www.pacer.gov/documents/pacermanual.pdf.
221. See supra Part IV.B.2.a–b.
222. See, e.g., Matt Blaze et al., Decentralized Trust Management, in 1996 IEEE SYMPOSIUM
ON SECURITY AND PRIVACY, at 164–73.
2012] ONLINE COURT RECORDS 823

formation is used to restrict access to records. 223 In addition, PACER


does not allow users to state their relationship to the data subject or to
provide evidence to support those claims, in contrast with courthouse
access where presenting and verifying such credentials are possible.
Thus, at this point in time, certain flows of information that are possi-
ble at physical courthouses are not possible online (at least, in the sys-
tems we studied).
5. Format of records. The format of the records affects the cost of
subsequent use and the dissemination of information in the records.
For example, searching a native PDF document224 is significantly less
costly than searching a scanned handwritten record. Our online
searches often yielded portable documents for which the first copy
was available for a nominal fee; 225 once obtained, these documents
could be subsequently used and disseminated without incurring any
additional cost. In contrast, in some physical courthouses, only paper
documents were available. 226 Making additional copies, at five cents
per page, increases the cost. Although one could disseminate
scanned paper copies without incurring additional cost, the resulting
documents could only be searched at significantly higher cost than
native PDF documents. 227
6. Human factors. Finally, human factors affect the cost of retriev-
ing information from court records. In particular, the user-interface
for searching records affects the cost in at least two ways. First, much
work in search engines has gone into figuring out how to present re-
sults so that users can quickly identify the results of greatest interest.
In Google Scholar, the feature of displaying snippets of text in addi-
tion to the title of returned result was particularly helpful. In con-
trast, searching at the physical courthouse involved using dated sys-

223. In one of the searches we tried, we were denied access to the documents related to
the court record, including the complaint form, summons orders, answer to the complaint
form, and the Social Security transcript filed. We could not tell from the information pro-
vided by PACER why access was denied to this particular court record. Further research
on this particular case suggests it may have been related to privacy concerns over medical
records or Social Security information or both.
224. A native PDF is one that is digitally converted from an electronic text format into
the PDF format. While converted in this manner, the text of the document may be
searched for specific terms and specific passages may be copied and pasted out of the PDF.
225. See supra Part IV.B.1.a. There is no fee for Google Scholar, but PACER charges a
fee for documents retrieved.
226. See supra Part IV.B.2.a.
227. Unlike native PDFs, scanned PDF documents are created from the scanned image
rather than the original metadata. Thus, the ability to search for specific words or phrases
is lost. Some software, which is capable of recognizing individual letters, can be used to
maintain the ability to search the document for specific words or phrases.
824 MARYLAND LAW REVIEW [Vol. 71:772

tems with poorly designed interfaces and interacting directly with


clerks. Second, a human interface in the clerk’s office provides a level
of informal social protection; 228 a user might be hesitant to ask for in-
formation that would indicate improper use of court records (for ex-
ample, a juror accessing a case file), but a user who is searching ano-
nymously online is not similarly inhibited.

V. EVALUATING ACCESS PRACTICES AND POLICIES


In Part IV, we revealed how the flows of information in court
records, shaped by explicit policy that has been enacted through the
media of hardcopy and local access terminals, may systematically be
altered by the adoption of Web-based publication and access. 229 Our
analysis makes clear, however, that the precise contours of the altera-
tions in flow are not determined by the choice of medium alone; ra-
ther, several specific further choices are crucial to the outcome, such
as search interfaces, indexing characteristics, and linkages to other in-
formation sources. 230 Since our inquiry is normative—what courts
ought to do—it must look beyond merely revealing differences to eva-
luating them. Guided by the framework of contextual integrity, as
outlined in Part III, this involves a two-layer evaluation: (a) taking
stock of the impacts of respective information flows in general moral
and political terms, and (b) establishing the significance of these im-
pacts in terms of values and purposes internal to the justice system
and courts, specifically. We follow in Part VI with recommendations
drawn from our two-layer analysis.

A. Impacts of Information Flows in General Moral and Political Terms


There are various sources from which to draw insight into the
impacts of information flows into and out of court records. For ex-
ample, the reasoning of judges as they decide whether to seal parts of
a record often reveals their assessments of impacts, as do rules and
policies explicitly adopted by respective courts. Finally, insights
drawn from privacy scholarship, in general, may inform our assess-
ment of information practices followed by courts.

228. See Winn, supra note 95, at 157 (“Perhaps the most significant change to the judi-
cial information ecosystem was the elimination of a human interface in the clerk’s office,
ending the informal social protections which formerly existed to control access to the case
file.”).
229. See supra Part IV.B.1.
230. See supra Part IV.C.
2012] ONLINE COURT RECORDS 825

While judges have cited shame, stigma, the negative judgment of


their communities, and chilling effects as reasons for granting ano-
nymity to rape victims, rape is not the only context in which such
chilling effects might come into play. A different type of chilling ar-
gument is found in In re Twentieth Century Fox Film Corp., 231 where a
New York court approved the request of Macaulay Culkin and the film
studio to seal a court-approved performance contract so as not to dis-
courage others from entering into contracts with children; the judge
also determined that the public’s right to know these details was
weak. 232 Adverse publicity is not in itself sufficient; for example, in In
re Azabu Buildings Co., 233 which involved a debtor and mortgage com-
pany, a federal court held that the threat of negative publicity was not
a compelling reason to seal a record. 234 In a case involving Bob Dylan,
a judge closed portions of the record deemed immaterial to the case
both to prevent unfair use of it by the plaintiff and to protect Dylan’s
privacy and reputation. 235
The unfair advantage one side may gain over the other has, on
other occasions, too, been found sufficient to restrict access to parts
of a record. 236 To protect the identity of an informant in an ongoing
investigation and assure the safety of her family, government requests
to seal records have been honored. 237 And although medical infor-
mation might be sealed or redacted if it is deemed stigmatizing and
immaterial to the case, judges have refused to do the same in cases
where it is material, such as personal injury litigation. 238

231. 190 A.D.2d 483 (N.Y. App. Div. 1993) (per curiam).
232. Id. at 486–87.
233. No. 05-50011, 2007 WL 461300 (Bankr. D. Haw. Feb. 7, 2007).
234. Id. at *2.
235. Damiano v. Sony Music Entm’t Inc., 168 F.R.D. 485, 491–93 (D.N.J. 1996) (finding
that the material was unrelated to public safety and that Dylan’s personal business affairs
are not a legitimate public concern, and suggesting that fairness and efficiency are pro-
moted among litigants as long as they share these materials with each other, and that this
is not contingent on public accessibility).
236. Cf. In re Twentieth Century Fox Film Corp., 190 A.D.2d at 488 (noting that Twentieth
Century Fox’s “relationship with its competitors, as well as with other artists in its employ,
could be compromised by the disclosure of the details of the contracts, which include in-
formation as to how it has marketed the subject motion picture”).
237. See, e.g., United States v. McVeigh, 918 F. Supp. 1452, 1458, 1466 (W.D. Okla. 1996)
(noting that statutory laws allows records to be sealed when they “‘could reasonably be ex-
pected to endanger the life or physical safety of any individual’” (quoting 5 U.S.C.
§ 552(b)(7))).
238. See, e.g., White v. Worthington Indus., Inc. Long Term Disability Income Plan, 266
F.R.D. 178, 196 (S.D. Ohio 2010) (requiring redacted version of record be publicly filed
because of its importance to the court’s decision and denying motion to file record under
seal).
826 MARYLAND LAW REVIEW [Vol. 71:772

Although courts handle access restrictions in a variety of ways,


there is substantial, if not universal, consistency surrounding certain
types of information, including trade secrets, other confidential
commercial research, national security concerns, and wholly private
family matters such as child custody or adoption. 239 Information that
may promote scandal, defamation, or unnecessary embarrassment is
handled similarly. 240 Information that “poses a serious threat of ha-
rassment, exploitation, physical intrusion . . . or the potential for
harm to third persons not parties to the litigation” also may be gener-
ally protected. 241 Certain categories of people, such as minors, 242 vic-
tims of rape and abuse, 243 and celebrities, 244 are also generally given
special consideration.
We do not mean to set much store by the particulars of any one
of these cases, except to illustrate the types of considerations that have
been and might be brought to bear in determining limitations on
access to information in a court record. To summarize: courts have
weighed the commitment to public access to records against consid-
erations of safety, stigma, shame, unfair disadvantage, and reputation-
al damage to concerned parties. At times, these factors are consi-
dered on a case-by-case basis and at times systematically through
principles and rules developed and adopted by court administrators.
It is worth noting that there are several ways to restrict information
flows into and out of records: redaction, sealing, and selective disclo-
sure to specific personnel.
While enthusiastic supporters of a transformation from local to
online access admit that limited but legitimate reasons for restricting
access exist, they deny that the medium or mode of access bears a sys-

239. See supra Part II.C; see also Doe v. Blue Cross & Blue Shield United, 112 F.3d 869,
872 (7th Cir. 1997) (stating that “[r]ecords or parts of records are sometimes sealed for
good reasons, including the protection of state secrets, trade secrets, and informers”); Hol-
land v. Eads, 614 So.2d 1012, 1016 (Ala. 1993) (listing scenarios where most courts pre-
sume that the court documents should remain sealed).
240. See, e.g., James v. Jacobson, 6 F.3d 233, 238–39 (4th Cir. 1993) (noting that “privacy
or confidentiality concerns are sometimes sufficiently critical that parties or witnesses
should be allowed” anonymity, and collecting cases).
241. Holland, 614 So.2d at 1016.
242. See, e.g., N.J. R. 1.38-3(d)(17) (protecting records of hearings on welfare or status
of child from public access); Blue Cross & Blue Shield United, 112 F.3d at 872 (“[F]ictitious
names are allowed when necessary to protect the privacy of children, rape victims, and
other particularly vulnerable parties or witnesses.”).
243. See, e.g., N.J. R. 1.38-3(d)(9)–(10) (protecting records related to domestic violence
and sexual offenses).
244. See generally John Gibeaut, Celebrity Justice: The Rich and Famous Get Star Treatment,
Creating the Appearance of a Two-Tiered Court System, 91 A.B.A.J. 43, 45–46 (2005) (discussing
how trial courts have protected celebrities’ information with increasing frequency).
2012] ONLINE COURT RECORDS 827

tematic relationship to these restrictions. Arminda Bepko, for exam-


ple, in pressing for online access to all records currently available in
courthouses, “argues that the constitutional and common law pre-
sumption in favor of public access to court documents should not
shift depending on the medium.” 245 Other advocates of unfettered
Internet access offer a more pragmatic rationale. Associate Justice
Albin of the Supreme Court of New Jersey, for example, a supporter
of greater public access to court records through the Internet, states:
“The information genie already has been released from the lamp, and
we cannot return to a simpler time when court records, although
open to the public, were stored in the practical obscurity of the clerk’s
office in the county courthouse.” 246
Given the widespread agreement that the choice of medium
makes a difference to degree of access, why is there disagreement
over policies that should govern publication and access? To answer
this question, it is useful to situate court records within the landscape
of government records more generally.
Government agencies amass vast quantities of information in
countless databases, making them available to citizens and other resi-
dents in highly variable arrangements. At one extreme, for records
such as geological surveys, FDA drug testing, and macroeconomic in-
dicators, unrestricted public access seems unquestionably warranted.
Let us call these “green records.” At the other extreme, for records
pertaining to “top secret” national security, law enforcement, and in-
dividual income tax, no access is given beyond a handful of autho-
rized personnel. Let us call these “red records.” In between, lies a
huge range of what we will call “orange records,” which are publicly
accessible with varying restrictions that may foreclose scrutiny of all or
parts of the records. In the case of the Census, for example, while
access to raw data is highly restricted, 247 public access is freely given to
anonymized aggregations. 248 Restrictions might be imposed through

245. Arminda Bradford Bepko, Note, Public Availability or Practical Obscurity: The Debate
Over Public Access to Court Records on the Internet, 49 N.Y.L. SCH. L. REV. 967, 968 (2005); see
also David Robinson et al., Government Data and the Invisible Hand, 11 YALE J.L. & TECH. 160,
166–67 (2009) (suggesting government open access to its data for public use).
246. REPORT ON PUBLIC ACCESS, supra note 15, at ii.
247. See 13 U.S.C. § 9(a) (2000) (restricting use, publication, and access of individual
data collected for the Census); Douglas Kysar, Kids & Cul-de-Sacs: Census 2000 and the Re-
production of Consumer Culture, 87 CORNELL L. REV. 853, 870 (2002) (reviewing U.S. CENSUS
BUREAU, U.S. DEP’T OF COMMERCE, CENSUS 2000: CENSUS OF POPULATION AND HOUSING
(2001)) (“[T]he Bureau has maintained a policy of strict confidentiality with respect to
individual census returns since at least 1840.”).
248. See 13 U.S.C. § 8(b) (2000) (stating that “tabulations and other statistical materials”
may be obtained); U.S. CENSUS BUREAU, DATA PROTECTION, http://www.census.gov/pri-
828 MARYLAND LAW REVIEW [Vol. 71:772

conditions on access, such as credentials or authorization, or by im-


posing limits on use. Court and DMV records (particularly following
the Drivers Privacy Protection Act) are also examples of orange
records. 249
When considering how to adapt existing policies to the new me-
dia of digital file storage and network access, the key challenges might
seem to be economic and engineering. For green records, for exam-
ple, this seems straightforwardly to be the case, namely, how to design
a new system that provides the most effective, efficient, and useful
access. As media for information delivery have progressed, and, with
the advent of digital electronic information technologies, improved
by orders of magnitude, it is reasonable to expect that socio-technical
systems for providing access will improve accordingly. The only coun-
tervailing consideration might be expense, as governing authorities
juggle diverse claims on limited resources. Because less-than-optimal
access increases the generalized costs (as defined in Part IV) of access
for citizens, resisting improvements might easily be read as opposition
to government openness.
Orange records—those governed by varying ensembles of rules—
present a greater challenge. The records administrator must engi-
neer access constraints not only to embody the ensemble of rules ac-
curately but—here we state the obvious—to adapt them appropriately
to characteristics of the medium; thus, the proverbial thick black
marker may be effective for selectively redacting print on paper;
locked filing cabinets effective for providing selective access to those
with keys; and an obliging, non-intrusive clerk effective for approx-
imating open access to information held in manila folders stored in
filing cabinets. When the recording medium changes, and the task of
implementing policies are revisited, indeterminacies may be encoun-
tered when a new medium allows new options or shuts down existing
ones. To give a simple example, redaction in electronic media could
be modeled in a variety of ways. 250 Although permanent erasure of
words within a document might most closely mimic the action of an
indelible black marker, an administrator, confronted with a new
technology allowing for reversible obfuscation, might find explicit
policy silent on the question of permanence or reversibility. New op-

vacy/data_protection/ (last revised Feb. 16, 2012) (explaining the obligation of Census
Bureau to provide statistics for public and private use).
249. Drivers Privacy Protection Act, 18 U.S.C. § 2721 (West Supp. 2011) (limiting ability
of state department of motor vehicles to disclose information collected in connection to
motor vehicle exception, subject to some exceptions).
250. See Lee, supra note 42.
2012] ONLINE COURT RECORDS 829

tions can be liberating, but they also reveal incompleteness in explicit


rules in relation to actions that simply were not possible before; what
might seem at first to be merely an engineering question turns out,
after all, to be a policy question. 251 Although some of the new options
opened by reengineering are neutral with respect to policy, others
might render policy ambiguous or, worse, may actually undermine
unarticulated policy expectations. It is important for policymakers
and administrators at least to recognize these possibilities and be
ready to make determinations, where necessary.
If the purpose of Part IV was to expose how a change in technical
systems may result in altered flows and novel access options, this Part’s
purpose is to expose where these options reveal policy ambiguities
and even undermine unexpressed policy expectations, particularly
those relating to privacy interests. Because migrating to the new me-
dium results in a lifting of constraints that had “naturally” been em-
bodied in prior media, it exposes a need to locate resulting flows that
contravene policy or expectation. Some of the trouble surrounding
Social Security numbers illustrates the need to be alert to inadvertent
gaps opened by technological choices. Against the backdrop of wide-
ly available personal information, Social Security numbers emerged as
valuable keys to identity theft. 252 To counteract the lifting of natural
barriers inherent in prior dominant media, as well as novel threats
posed by networked data sources, lawmakers and regulators have ac-
knowledged the need to limit the availability of Social Security num-
bers, both retroactively and proactively, in court and other public
records. 253
In light of transformations in flow due to the lowered costs of
access, 254 let us start by considering the criteria that courts—including
judges and administrators—have consistently recognized as legitimate
for restricting access to records or parts of records, whether in regula-
tions or in case-by-case assessments. Whatever the case was favoring
restricting access (for example, by sealing) to specific types of infor-
mation, such as financial, medical, and adoption, it is surely streng-
thened by the prospect of radically lowered costs of access and ought
to add significant weight in case-by-case balancing of potential harms

251. See, e.g., LAWRENCE LESSIG, CODE AND OTHER LAWS OF CYBERSPACE 111–18 (1999)
(discussing the development of wiretapping technology and the Court’s approach to its
use under the Fourth Amendment).
252. See Bepko, supra note 245, at 998 (suggesting that Social Security numbers stolen
from court documents can be used to commit fraud).
253. See, e.g., N.J. R. 1:38-7(a)–(b) (prohibiting the disclosure Social Security numbers
and other personal identifiers from any court document or pleading).
254. See supra Part IV.
830 MARYLAND LAW REVIEW [Vol. 71:772

of access against benefits. An argument may also be made, however,


for imposing restrictions beyond those criteria readily acknowledged
by judges and court administrators. 255 Powerful new capacities to in-
fer financial, family, or medical information, for example, from com-
binations of other more innocuous information types suggest a need
to expand the class of potentially stigmatizing information that can
justify sealing. 256 Under the “Mosaic Theory,” an equivalent argument
is offered to justify limits on access to a wider swathe of information in
government hands on grounds of potential threats to national securi-
ty. 257 Even information not previously thought to be risky, by itself,
can become risky when linked to additional publicly available infor-
mation. 258
Out of concern for the intimately personal and sometimes even
embarrassing information revealed in many cases, courts have been
prepared to seal or redact information in records that could cause
scandal, defamation, harassment, ridicule, or unnecessary attention
and embarrassment, giving special consideration to minors, victims of
rape and domestic abuse, third persons not party to the litigation,
and, in some instances, celebrities. The point highlighted here is
that given the magnitude of impact when information is posted on-
line, the chances of scandal, defamation, embarrassment, and unne-
cessary attention are increased significantly, particularly when nosy
neighbors, celebrity followers, or anyone with an interest in a particu-
lar individual is able to search online and, at low cost, associate that
individual with information from a court record. An extreme exam-
ple of the impact of linking publicly available information is a case in
France, where Eric Schmidt, the former CEO of Google, was found
guilty of defamation because Google Suggest brought up terms such
as “Satanist,” “Rape,” “prison,” and “rapist” when the name of a man
identified in court records only as Mr. X was entered into a search
box. 259 Mr. X had been convicted of “corruption of a minor” in a

255. For example, the Judicial Conference Committee on the Court Administration and
Case Management adopted a set of guidelines that require sensitive information to be re-
dacted in civil and bankruptcy cases. Bepko, supra note 245, at 977. Remote electronic
access is also denied to any documents in criminal cases. Id.
256. David E. Pozen, Note, The Mosaic Theory, National Security, and the Freedom of Informa-
tion Act, 115 YALE L.J. 628, 630 (2005).
257. Id. at 645–47 (noting that the mosaic theory has lead to heightened protection of
national security information by the Executive Branch, including several new limits to the
Freedom of Information Act).
258. Id. at 630 (“In the context of national security, the mosaic theory suggests the po-
tential for an adversary to deduce from independently innocuous facts a strategic vulnera-
bility, exploitable for malevolent ends.”).
259. Weintraub, supra note 197.
2012] ONLINE COURT RECORDS 831

French court. 260 His conviction triggered the terms returned by


Google Suggest, but the associated terms were clearly far more damn-
ing than the charges leveled against the man, who upon appeal was
given merely a three-year suspended sentence and a fine. 261 A more
typical case, however, might involve web searches that potentially re-
veal involvement in legal action. 262
It may be difficult to sympathize with criminals or those who have
deservedly been sued wishing to avoid scandal and embarrassment.
There are, however, many non-criminals identified in court records
who are subject to the same treatment, including persons found in-
nocent after having been arrested and charged, those who have sued
with justification, those implicated as third-parties, and witnesses, to
name but a few. It is worth weighing the added burden to these indi-
viduals, at least those who may suffer embarrassment, ridicule, pre-
judicial discrimination in seeking employment and housing, and even
simply uninvited attention as a result of linkable information.
The potential for harassment, however, deserves special mention.
There have been chilling stories of people suffering grossly dispropor-
tionate harassment for misdemeanors that happen to have captured
the attention of the online masses, such as the infamous “dog poop
girl” in Korea. 263 In China, the Internet has also facilitated so-called
“human flesh search engines,” a form of vigilantism aimed at punish-
ing accused wrongdoers through coordinated, massive online and of-
fline reaction. 264 While a single reproach, deserved or not, can be
unpleasant, the same reproach coming from thousands is surely ha-
rassment that no civilized society should encourage. The capacity of
the Internet to unleash collective retribution in this manner deserves
to be among the factors considered when balanced against the post-
ing online of personal information in a court record. This, along with
other considerations suggests that with higher stakes, the bar for re-
stricting access should be lowered.

260. Id.
261. Id.
262. See, e.g., David Kravets, Mug Shot Industry Will Dig Up Your Past, Charge You to Bury It
Again, WIRED.COM (Aug. 2, 2011, 1:52 PM), http://www.wired.com/threatlevel/
2011/08/mugshots/.
263. See Jonathan Krim, Subway Fracas Escalates Into Test of the Internet’s Power to Shame,
WASH. POST, July 7, 2005, at D1 (describing the impact of blogs in forcing a South Korean
student to quit college after she refused to clean up after her dog on a subway, pictures
surfaced of her online, and people discovered her identity and vilified her in public).
264. Ariana Eunjung Cha & Jill Drew, New Freedom, and Peril, in Online Criticism of China,
WASH. POST, Apr. 17, 2008, at A1.
832 MARYLAND LAW REVIEW [Vol. 71:772

B. Values and Purposes Internal to Courts and the Justice System


Part V.A addressed the first layer of evaluation prescribed by the
framework of contextual integrity, taking stock of impacts of new pat-
terns of information flow with ethical import. This section addresses
the second layer of evaluation by considering the significance of these
impacts for the advancement of context-specific values, ends, and
purposes, particularly as compared with prior practice. In the cases of
the U.S. Census Bureau or Internal Revenue Service, for example,
both known for highly restrictive access policies for raw data, the im-
pacts of more relaxed privacy policies might include harms to indi-
viduals, an unbalancing of power relations between citizens and gov-
ernment, and so forth. Yet, in both cases, an argument that has held
sway highlights deleterious consequences for the effective functioning
of respective agencies. Because individuals will be less likely to partic-
ipate and less likely to do so honestly without appropriate restrictions
on flows, the agencies’ goals of universal and accurate reach would be
undermined. 265
So, too, must we investigate how changes to constraints on access
to courts records, including constraints motivated by privacy con-
cerns, affect not only individual data subjects but also the degree of
success courts may have in advancing context-specific goals and val-
ues. What precisely are the goals and values of courts within the jus-
tice systems of liberal democracies is likely to be debated and dis-
cussed, even in well-ordered, free societies, and this Article is no place
for a comprehensive account of these. Instead, we draw from a set of
goals and values that are widely cited, solidly robust, and relatively un-
controversial, which can be found in scholarly articles and formal and
informal accounts of individual commentators and professional or-
ganizations. These serve as a basis for our argument, which proceeds
as follows: brief discussions of (1) ends and purposes (or goals); (2)
values served by courts; (3) how court records function to promote
goals and values; and (4) how deployment of networked access that
disrupts information flows may undermine the attainment of these
goals and values.

1. Ends and Purposes


Most would agree that courts serve to establish that a crime has
occurred and pass judgment on criminal guilt. They adjudicate, or
resolve conflicts and disputes among public and private parties in ac-

265. IRS, U.S DEP’T OF THE TREASURY, DISCLOSURE & PRIVACY LAW REFERENCE GUIDE 1-1
to 1-9 (2007).
2012] ONLINE COURT RECORDS 833

cordance with prevailing law. They officiate key societal relationships,


such as marriage and divorce, and establish child custody. Courts are
empowered not only to assess crime and adjudicate disputes but also
to attribute blame, assess liability, and determine punishment, reme-
dies, compensation, fines, and injunctive relief. According to the Na-
tional Association of Court Management:
Courts reinforce the authority of the state and the legitimate
use of force and protect individuals against the arbitrary use
of governmental power.
...
. . . Courts exist to do justice, to guarantee liberty, to en-
hance social order, to resolve disputes, to maintain rule of
law, to provide for equal protection, and to ensure due
process of law. 266
The association further states:
Only the judiciary can definitively determine who is to pre-
vail in the inevitable conflicts that arise between individuals;
between government and the governed, including those ac-
cused by the state of violating the law; between individuals
and corporations; and between organizations, both public
and private. . . . They resolve disputes by applying the law to
the facts of particular cases independently and impartially. 267
Similarly, in Sharon Rodrick’s words, “Judicial power is con-
cerned with the determination of disputes and the making of orders
concerning the existing rights, duties and liabilities of persons in-
volved in proceedings before the courts.” 268 Courts are not merely
applying law but, more often than not, they are creating law in their
interpretations and judgments and in setting precedents for future
courts. Within these broad categories, there are, of course, myriad ac-
tivities and practices that support them.

2. Values
As one of the pillars of the tripartite system of democratic gover-
nance, courts absorb core values of the overarching political system
while paying special heed to values that are tied to their specific

266. NATIONAL ASS’N FOR COURT MGMT., CORE COMPETENCY CURRICULUM GUIDELINES:
PURPOSES AND RESPONSIBILITIES OF COURTS, 1 (2003) [hereinafter PURPOSES], available at
http://www.nacmnet.org/sites/default/files/images/1purposes_0.pdf.
267. Id.
268. Sharon Rodrick, Open Justice and Suppressing Evidence of Police Methods: The Positions
in Canada and Australia (Part One), 31 MELB. U. L. REV. 171, 185 (2007).
834 MARYLAND LAW REVIEW [Vol. 71:772

mandate. 269 In listing the courts’ purposes as adjudicating disputes,


establishing guilt, meting out punishment, and so forth, it is essential
to pair these with the values to which courts subscribe: values of jus-
tice, fairness, proportionality, impartiality, predictability, freedom
from bias (“equal protection”), independence from inappropriate in-
fluence, and due process. 270 As a wing of liberal democracy, sup-
ported by and serving citizens, courts strive for effectiveness, efficien-
cy, competence, trustworthiness, accountability, service excellence,
openness, and transparency. 271 Courts must not only administer jus-
tice but also do so with the appearance of justice. Rodrick writes:
Much has been written about the purpose and value of open
justice. Of prime importance is the belief that open justice
enhances the integrity, accountability and performance of
those who are involved in the administration of justice. For
example, it is supposed that openness makes judges more
accountable for the manner in which they exercise the judi-
cial power that is vested in them; secret courts are regarded
as having a propensity to spawn corruption. Witnesses are
thought to be more likely to tell the truth if they have to tes-
tify orally in open court. For their part, the public, having
observed responsible and truthful behaviour on the part of
the judges and witnesses, will have increased confidence in
the operation of the courts and a greater understanding of
society’s laws and legal system. Any perceived shortcomings
in the behaviour of particular individuals or in the substance
of the law or its application in a particular case can be pub-
licly scrutinised and, perhaps, corrected. 272
One final remark on ends, purposes, and values: courts might do
all of the above and yet still fall short of their responsibilities to socie-
ty. It is not sufficient for courts, if employed by members of society, to
mete out justice and perform their tasks fairly, transparently, and so
forth, if they are under-utilized. Although there are means outside of
the courts to deal with bad actors and alternative institutions for re-

269. See John F. Manning, Separation of Powers as Ordinary Interpretation, 124 HARV. L. REV.
1939, 1942–43 (2011) (recognizing two approaches to judicial separation of powers: (1) a
functionalist approach that eschews the separation of powers notion; and (2) a formalist
approach in which courts are bound by the U.S. Constitution).
270. See PURPOSES, supra note 266, at 1–2.
271. Natalie Gomez-Velez, Internet Access to Court Records—Balancing Public Access and Pri-
vacy, 51 LOY. L. REV. 365, 369 (2005).
272. Rodrick, supra note 268, at 172 (footnote omitted).
2012] ONLINE COURT RECORDS 835

solving disputes, courts carry the authority of the law, must act in ac-
cordance with the law, and often create law when they act.

3. Function of Court Records


Internal procedures: Records are inherent to the internal function-
ing of courts as a court of record. 273 Decisions made in court must be
noted so they can be properly acted upon: perpetrators fined or im-
prisoned, debts paid, harms compensated, and so forth. Records not
only document these decisions, they convey them to other collabora-
tors within the justice system, such as prisons and law enforcement. 274
Court records constitute a repository of information necessary first for
the processing of cases at the trial level, and later as vehicles for
transmitting key facts, as cases are appealed. Good records systems
can contribute to efficiency, integrity, and fidelity of court function.
Educate and inform: Records contribute to the quality of courts’
functioning by educating and informing those who work in or with, or
are training to work in or with, the legal system, including lawyers,
judges, legislators, court administrators, and academics. 275 Records
help make these crucial participants better at their jobs. Although
most legal professionals pay for access to records from the dominant
third party services that organize records and include citations, such
as Westlaw and Lexis, full records remain freely available. By com-
municating process, reasoning, and precedent, court records pro-
mote competence, knowledge, consistency, and understanding.
For society at large, court records provide a window into how
courts function. 276 Though most likely interpreted by the media and
not read directly, court records facilitate for the public a better un-
derstanding of the judicial branch of government, specifically, the
court system, how and when to use it, and what to expect if one
does. 277 Court records serve the end of familiarizing citizens with the

273. Interview with Jim Rebo, Chief Security Officer, New Jersey Courts, in Trenton, N.J.
(May 15, 2009).
274. Margo Schlanger & Denise Lieberman, Using Court Records for Research, Teaching,
and Policymaking: The Civil Rights Litigation Clearinghouse, 75 UMKC L. REV. 155, 160 (2006).
275. Id. at 162.
276. See Gomez-Velez, supra note 271, at 402–03 (“[A]n important counterpoint to con-
cerns about individual privacy in the context of placing court records on the Internet is
the competing and legally recognized interest in government transparency—in providing
public access to information about government functions.”).
277. See id. at 406 n.136 (“‘Electronic access ensures that the public and the news media
can oversee how justice is administered like never before.’” (quoting REPORTERS COMM.
FOR FREEDOM OF THE PRESS, ELECTRONIC ACCESS TO COURT RECORDS: ENSURING ACCESS
IN THE PUBLIC INTEREST)).
836 MARYLAND LAW REVIEW [Vol. 71:772

workings of government; an educated and informed citizenry contri-


butes to better government in all branches. 278
Public oversight, open justice, and transparency: As windows to the
courts, court records serve not only the educational functions noted
above, which contribute to effective functioning, but also serve as in-
ternal and external checks. Whether via direct access or via interpre-
tations by NGOs and the media, court records hold judges and other
court officials accountable for the quality of their work. Calls for gov-
ernment transparency and accountability are, to a large extent, ans-
wered by the free availability of records, the right of any member of
society, for whatever reason, to inspect records so they may judge for
themselves whether the courts are functioning as they should, serving
the public effectively (and cost effectively) and impartially. Transpa-
rency and openness of courts to the citizens of a democracy are as vi-
tal to this branch of government as to others, whether to educate, sa-
tisfy public curiosity, ensure oversight and accountability, protect
against corruption and abuses of government power, or provide
grounds for the appointment or reappointment, or election or reelec-
tion, of judges and other court officials.

4. Disruptive Information Flows


In scholarly discussions of online access, deliberations commis-
sioned by courts themselves, 279 and advocacy rhetoric of public inter-
est organizations, the rationale most consistently offered in favor of
boosted access is transparency of court functioning to the public. 280
Our analysis, in terms of the function courts records serve, leads to a
similar conclusion, for most of the other functions can be served just
as well by restricting access to authorized personnel only—such as
court and prison officials, judges, authorized social workers, lawyers,
and bail bondsmen—or by heavily limiting what parts of records are
disseminated. When judges and court administrators weigh the vari-
ous and possibly conflicting interests for and against placing docu-
ments or information into the public court record, they cite most fre-
quently this right of the governed to inspect, take stock of, and know
how they are being governed and how the institutions of their gov-
ernment are functioning.

278. Id. at 402–03.


279. See, e.g., id. at 368 n.3 (noting that New York commissioned a study on making its
government documents more accessible to the public).
280. See, e.g., Schwartz, supra note 14 (chronicling an entrepreneur attempting to start a
database with free access to court documents under the auspice of openness and transpa-
rency).
2012] ONLINE COURT RECORDS 837

Part V.A argued that, when balancing and trading off interests
against one another in order to formulate explicit rules for so-called
“orange” records, properties of underlying distribution media cannot
be ignored because they may well make material differences to these
interests. Following this logic, we revealed ways that online access can
increase the hardship of participants in court cases. Prima facie con-
clusions drawn on grounds of this observation need to be considered
in light of the second layer of analysis prescribed by the theory of con-
textual integrity. 281
The array of affected interests, although an essential part, is not
the whole story. How significant the disruptions are to the further-
ance of context-specific ends, purposes, and values is the rest of it.
Rodrick voices a similar concern in the Australian context:
[C]ourts have taken the view that the principle of open jus-
tice is so fundamental that it can be curtailed only when ne-
cessary in the interests of the administration of justice in the
particular proceeding. . . . [T]he proceeding could not ef-
fectively continue (or would be prejudiced or frustrated in
some real and tangible way), or the decision of the court
would be deprived of practical utility. 282
Further, “[t]his is because open justice, which is primarily valued for
its contribution to the administration of justice, must yield to the
need to secure the administration of justice in the unusual event that
publicity would be to its detriment.” 283
In other words, when evaluating the consequences of posting
records online, although threats of harm posed to information sub-
jects are important, our analysis highlights another set of key conse-
quences, in particular, “the administration of justice” itself, that is,
goals and values of courts within the judicial system. If, say, victims of
certain types of crimes or wrongs do not bring them to the courts or
choose alternative venues to settle them, in order to avoid the addi-
tional burdens of publicity, there may be cause for concern. 284 Perpe-
trators may go unpunished and suffering uncompensated as the

281. See Adam Barth, Anupam Datta, John C. Mitchell & Helen Nissenbaum, Privacy and
Contextual Integrity: Framework and Applications, in 2006 IEEE SYMPOSIUM ON SECURITY AND
PRIVACY 184, 184 (“Contextual integrity is a conceptual framework for understanding pri-
vacy expectations and their implications developed in the literature on law, public policy,
and political philosophy.”).
282. Rodrick, supra note 268, at 185–86 (footnote omitted).
283. Id. at 187.
284. See, e.g., Strahilevitz, supra note 53, at 1247–48, 1252 (noting that people use ano-
nymous forums online as cheaper alternatives to litigation).
838 MARYLAND LAW REVIEW [Vol. 71:772

court’s role is diminished in areas of social life that might benefit


from the justice system’s intervention. Although there may be good
reasons to choose, say, mediation or arbitration, society may lose out
on the development of important legal precedent and remain igno-
rant of worrying social trends. 285 Further, the courts are important
constituents of democratic governance and are accountable to the
public in ways that other systems of mediation and arbitration may
not be; it would be cause for concern if their range of influence were
to tilt or erode because parties were choosing these other systems,
possibly less exposed to public oversight.
Equally as important is the role of non-parties, who make essen-
tial contributions to court proceedings by acting as jurors, as wit-
nesses, or in other capacities. Grayson Barber and Peter Winn have
highlighted the plight of such “unrepresented” parties, including
third parties mentioned in court proceedings whose privacy interests
might simply be overlooked. 286 These oversights might be irrelevant
where court records are accessed via traditional paper methods. In a
world of open access to full and complete records, however, certain
indexing choices made by court administrators would allow a Google
search on any of the parties’ names to reveal their involvement in a
case. 287 Even a slight disincentive such as this might tip the scale
against offering one’s services to a court.
Two final points linking new patterns of disclosure with purposes
and values: the first concerns justice and proportionality. To the ex-
tent that the placement of records online gives rise to further hard-
ships, such as harassment, and discrimination in job and other oppor-
tunities, it raises questions about disproportionate punishment. One
may have little sympathy for serious criminals, particularly for those
who have sexually assaulted children, considering it their just deserts
and pointing out the high rates of recidivism. But these are most like-
ly a small percentage of those affected by, say, harassment.

285. See id. at 1258 (highlighting the potential loss of important precedent if a litigant
had not been allowed to pursue trial under a pseudonym).
286. See, e.g., Grayson Barber, Personal Information in Government Records: Protecting the
Public Interest in Privacy, 25 ST. LOUIS U. PUB. L. REV. 63, 63–67 (2006) (arguing the need
for courts to recognize their “special obligation to protect the public’s interest in individu-
al privacy” with respect to government records containing private information); Winn, su-
pra note 95, at 152 (“Because of the nature of the adversarial system, the interest of the
public in the transparency of judicial records and the interest of unrepresented third par-
ties in protecting sensitive private information are equally ignored as the parties before the
court pursue their own personal interests.”).
287. E.g., Marder, supra note 2, at 450 (describing how Proposition 8 opponents in Cali-
fornia used “Eightmaps.com” and Google Maps to locate the homes and email addresses
of Proposition 8 supporters and then harass and threaten them).
2012] ONLINE COURT RECORDS 839

The fate of expunction is a second related point. Many states of-


fer those found guilty of certain categories of offenses, duly punished
and clear of any further offenses, the opportunity to have their
records expunged. 288 The same is offered to persons who have been
arrested but against whom no further action was taken. Online post-
ing could nullify, or seriously limit, the chance of disassociating one’s
name from such convictions or arrests. 289

VI. RECOMMENDATIONS FOR ACCESS


In defining the aims of our research, we have sought to develop a
general line of reasoning for considering privacy while focusing on
the transition from locally accessed court records to networked elec-
tronic records. 290 Our arguments and findings, thus far, point to vari-
ous approaches that may be taken to address key concerns. 291 Given
the vastness of the enterprise and variation across courts and states,
the business of embodying a particular, principled approach in specif-
ic rules for specific courts will require substantial adaptation and in-
terpretation. Thus, for now, we limit our concluding discussion with
brief descriptions of alternatives we favor, as well as respective impli-
cations for the design of electronic court-record systems and man-
agement of associated information-handling practices. Although we
are able to present only open-ended accounts, we aim to provide
enough substance to stimulate meaningful debate.
• Option 1: A radical departure from the status quo
One global policy option that follows from our discussion is for
courts to “sanitize” records by redacting proper names and possibly
other immediately identifying information, such as Social Security
numbers, bank account numbers, or coded biometric identifiers of all
parties and other participants (for example, witnesses, jurors, etc.),
before releasing them to the public via the Web or local portals (for
example, local terminals or paper files). Given the possibility of on-

288. See, e.g., N.J. STAT. ANN. § 2C:52-2 (West 2005 & Supp. 2011) (providing that “a
person [who] has been convicted of a crime under the laws of this State and who has not
been convicted of any prior or subsequent crime . . . and has not been adjudged a disor-
derly person or petty disorderly person on more than two occasions may, after the expira-
tion of a period of 10 years from the date of his conviction, payment of fine, satisfactory
completion of probation or parole, or release from incarceration . . . present a duly veri-
fied petition . . . to the Superior Court in the county in which the conviction was entered
praying that such conviction and all records and information pertaining thereto be ex-
punged.”).
289. See, e.g., PURPOSES, supra note 266.
290. See supra Part III (defending a theory of information privacy).
291. See supra Part V (describing policy based on normative information flow analysis).
840 MARYLAND LAW REVIEW [Vol. 71:772

line posting and the potential to increase the burden on parties as


well as unrepresented participants, instead of relying on judges and
administrators to make determinations on a field-by-field, case-by-case
basis, this alternative would constitute a bright line policy across the
board. Its purpose would be to protect threats of harms to interested
parties as well as threats to the attainment of ends, purposes, 292 and
values 293 served by courts posed by the far-reaching disruptions of in-
formation flows, 294 discussed above.
Although, in the United States, it would constitute a radical de-
parture from the status quo to provide public access only to sanitized
records, 295 experience with Jane Roe and John Doe cases and the
practice in many European countries of maintaining strict anonymity
for the parties in court cases suggests it can be done. 296 Our main cri-
teria for evaluating this option, and the ones that follow, are drawn
from the approach we have developed, generally, for addressing wor-
rying implications for privacy of online placement. 297 This evaluation
considers advantages as well as costs and drawbacks.
Redacting immediately identifying information somewhat lowers
the stakes of providing free and unrestricted online access and dimi-
nishes pressure to seal or redact other information typically deemed
sensitive, such as medical and financial information. It protects
against Natalie Gomez-Velez’s worry that changes in dissemination
patterns due to online access of personal information in court records
will not necessarily serve the intended values, increasing such impro-
per uses “as identity theft, stalking, discrimination, locating domestic
violence victims, and interference with business and social relation-
ships, rather than for the appropriate oversight, educative, and ac-
countability reasons for which court records are made public.” 298 Ca-
ren Myers Morrison, in an article focusing on Internet access to

292. See supra Part V.B.1 (discussing the ends and purposes of the justice system).
293. See supra Part V.B.2 (discussing the values of the justice system).
294. See supra Part V.B.3 (discussing potential harm caused by information flows).
295. See Marder, supra note 2, at 453 (noting that currently neither courts nor lawyers
accept responsibility for redacting sensitive information).
296. See, e.g., Elena Larrauri, Conviction Records in Spain: Obstacles to Reintegration of Of-
fenders, 3 EUR. J. PROBATION 50, 51–52 n.11 (2011) (noting the Spanish practice of render-
ing anonymous decisions, i.e., not publishing names, except for Constitutional Court deci-
sions); see also James B. Jacobs & Elena Laurrauri, Are Criminal Convictions a Public Matter?
The USA and Spain, 14 PUNISHMENT & SOC’Y 1, 3–28 (Jan. 2012).
297. See supra Part V.A (arguing for a balancing approach to resolve competing interests
with respect to governing information flow).
298. Gomez-Velez, supra note 271, at 371.
2012] ONLINE COURT RECORDS 841

criminal justice records, questions the need for proper names in


records with a similar challenge:
What is the information of value that the public needs to
know? Does the public need to know that an individual in-
dicted for distributing five kilograms of cocaine, which
would ordinarily entail a mandatory minimum sentence of
ten years, cooperated with the government and received a
sentence of thirty-six months, or does it need to know that
Billy Costigan, in particular, cooperated with the govern-
ment? 299
Following this line of reasoning, sanitized records could offer a
good compromise between the respective currents of privacy and
open government, on the one hand, by limiting improper and harm-
ful uses of personal information in court records while, 300 on the oth-
er hand, continuing to support key functions of court records in the
attainment of core values and goals. 301 Quality of court and profes-
sional performance, education, public oversight, transparency, and
accountability 302 will not be diminished because the names of lawyers
and court and other officials of the justice system (such as judges)
would not be redacted under this approach. Meanwhile, the content
of actual value to the public—decisions, reasoning, and other mate-
rials necessary to evaluate the quality of a judgment 303—will be just as
available as before, indeed, more available than before, because they
would be posted online. In other words, sanitizing records before
posting them online promises protections against threats to privacy
without significantly compromising values and goals of the courts.
While this option embodies a clear and well-justified principle of
action, it remains vulnerable to criticism from at least two sources,
one technical and the other normative. On the technical front, the
ongoing quagmire surrounding anonymization—its limits and wheth-
er it is even possible—surely must raise doubts over the possibility that

299. Caren Myers Morrison, Privacy, Accountability and the Cooperating Defendant: Towards
a New Role for Internet Access to Court Records, 62 VAND. L. REV. 921, 971 (2009) (footnote
omitted). Of course, a litigant or defendant’s identity may in fact matter, for example if a
particular judge is alleged to give harsher sentences to defendants of a particular race, or if
the same plaintiff repeatedly brings harassing lawsuits.
300. See supra Part V.B.3 (discussing the moral and political impacts of information
flows).
301. See supra Part V.B.3 (discussing the function of court records in relation to the key
values of the justice system).
302. See supra Part V.B.3 (discussing how court records serve these key functions).
303. See supra Part V.B.3 (explaining the importance to the public of this information
contained in court records).
842 MARYLAND LAW REVIEW [Vol. 71:772

removing proper names and other obvious identifiers would effective-


ly shield the identities of those involved in a case, either parties or
others, particularly if the seekers of information are armed with addi-
tional background knowledge. 304 We wonder, however, whether such
doubts are sufficient to foreclose this option. Although it may be poss-
ible for a highly motivated searcher to re-identify key individuals in a
case, redaction of proper names will still protect against serendipitous
discovery of someone’s involvement in a case as a result of an Internet
search on her name. Accordingly, the technical challenges posed by
anonymization, generally, must be evaluated in light of the specific
challenges of sanitizing court records. Imperfect as it is, simple redac-
tion may constitute the most practically feasible option not, perhaps,
compared with perfect anonymization but compared with doing noth-
ing at all. Going forward, therefore, we would like to see further dis-
cussion of this point considering not only, for example, what types of
information might need to be redacted from a record, or what fields
should and should not be indexed, in order to ensure that identifying
information be shielded, but also questions economic in nature, such
as, how much effort—or cost, as we have called it above—searchers or
search services would likely be willing to expend to re-identify indi-
vidual parties in a particular record of interest or to do the same for
records in bulk.
No matter how reasonable this option might be on principled
grounds, the break from tradition, the thoroughgoing changes in the
way cases proceed and the ways they are recorded and even named
make it an unlikely candidate in practice, not least because it may in-
fringe on media and citizens’ First Amendment rights by imposing a
form of prior restraints on access. Critics might argue that by redact-
ing names, society will lose an important dimension of answerability
from the courts, for example, assurances that there is no discrimina-
tion for or against plaintiffs or defendants based on race, ethnicity,
economic standing, or other inappropriate dimensions. This point is
important to consider, both to weigh its value against the benefits of
redaction and, in practical terms, to consider how important searcha-
ble names are to address it.

304. Linking attacks on relational databases have demonstrated this inadequacy. See,
e.g., Arvind Narayanan & Vitaly Shmatikov, Robust De-anonymization of Large Sparse Datasets,
in 2008 IEEE SYMPOSIUM ON SECURITY AND PRIVACY 111, 111 (noting how, even in databas-
es where names and Social Security numbers are redacted, individuals can nevertheless
“use background knowledge and cross-correlation with other databases to re-identify” the
anonymized individuals); Paul Ohm, Broken Promises of Privacy: Responding to the Surprising
Failure of Anonymization, 57 UCLA L. REV. 1701, 1703–04 (2010) (noting how “adversaries
can often reidentify or deanonymize the people hidden in an anonymized database”).
2012] ONLINE COURT RECORDS 843

Furthermore, the change in this aspect of court administration


would be costly, though already state courts have expended millions
of dollars in the transformation to electronic systems. 305 Besides the
general resistance one would expect from those for whom a change of
this scope would be simply too risky to consider, there are many par-
ties whose interests in fully identified open records would be placed at
risk, from reporters to information brokers.
Variations: A variant on this approach would sanitize the record
of identifying information of all those named in a case except the par-
ties themselves. This alternative would address some of the normative
concerns 306 and would diminish the burden on technical require-
ments, as non-party participants in a case, despite important contribu-
tions to the functioning of courts and the justice system, would almost
always be of far lesser interest to third parties.
• Option 2: A two-tiered alternative
A second option would retain the status quo for local access but
produce a sanitized version for the open, indexable Web. In other
words, two versions of court records would be produced, one available
to the public at local courthouses that includes all information cur-
rently approved and available for access, the other, posted online,
scrubbed of identifying information in the manner discussed above.
This would eliminate concerns about prior restraints, since public
access itself would not be curtailed—only online access would be af-
fected by this proposal. Because the First Amendment does not re-
quire courts to post their records online but only to make them ac-
cessible at the courthouse for inspection and copying, courts are free
to post limited or sanitized versions of court records online as long as
the full paper record is available at the courthouse.
While this option would allay major constitutional concerns, it
would nevertheless continue to support the existing practice of com-
mercial data aggregators paying employees to camp out at courthous-
es, manually copying records into their data repositories. Further-
more, it does not sufficiently address concerns expressed over
inappropriate flows of personal information from court records, and
does not stop the circulation of problematic information but rather

305. See, e.g., Megan Poinski, Electronic Court System Will Make Judiciary Paperless, Judge
Says, MARYLANDREPORTER.COM (Jan. 14, 2011), http://marylandreporter.com/2011/
01/14/electronic-court-system-will-make-judiciary-paperless-judge-says/ (discussing how
Maryland’s transformation to electronic filing will cost between $50 million and $60 mil-
lion).
306. See supra Part V.A.
844 MARYLAND LAW REVIEW [Vol. 71:772

results in uneven access to those who can afford to pay commercial


data aggregators for information.
A two-tiered approach that results in uneven access may, howev-
er, yield footholds for policy by holding identifiable third parties,
such as data aggregators, to higher standards of accountability. Such
policies might impose constraints on how the information flows and
uses to which it is put. Ideally, it would also hold third-party aggrega-
tors to higher quality standards for the data itself. 307
• Option 3: Fine-grained differential access
The broad policy options we have outlined thus far approximate
solutions that would take into consideration finer detail, such as the
roles of parties requesting access, the roles of information subjects,
the types of information, and the conditions under which information
is granted. This ideal is less complex than it may seem, as it genera-
lizes what is already common practice, for example, sealing informa-
tion about minors from public scrutiny that, presumably, is provided
to social workers, or redacting certain types of information from the
record that, ultimately, will be provided to the public. 308 Our argu-
ment has been that a reconsideration of common practice is necessary
in light of disruptions to information flows due to online access (dis-
cussed above). 309 But these same developments in information
science and media technologies also offer greatly enhanced capacities
for managing the additional complexity of finer-grain rules, as long as
the systems being built for use in courts take advantage of them. 310
In creating digital records, system designers have at their disposal
the means of building structure into them, including the ability to tag
data fields. This structure would allow constraints on selective access
to various parties based on rules governing differentiated access privi-
leges to differently tagged fields of information. Thus, for example, a
policy with a default to redact from a publicly accessible record the
names of all non-parties, including members of juries, witnesses, and
those inadvertently implicated through the case, which may have
been practically impossible in the past, becomes quite routine with

307. Interview with Grayson Barber, Grayson Barber LLC (April 2009).
308. See supra Part II.C (explaining the mechanisms and processes through which courts
decide whether or not to redact or seal information in their records).
309. See supra Part V.B.4 (noting the problems with online information flows and pro-
posing various prescriptions).
310. See Gomez-Velez, supra note 271, at 421–22 (noting those who have “argue[d] that
the access/privacy tension can be resolved through the use of technology, including the
use of computer programs to redact sensitive data elements or to anonymize informa-
tion”).
2012] ONLINE COURT RECORDS 845

the technical tools now available. Rule sets could be finely tuned to
surrounding conditions and, as prescribed by contextual integrity,
drawn from prior convention honed by context specific values and
goals. 311 This promising option would require that systems currently
being designed for state courts be fitted with necessary components.

VII. CONCLUSION: A HIDDEN VARIABLE


There are clearly powerful voices in favor of open court
records. 312 Despite this legacy, court records, according to our color-
coding scheme, have never been green but always orange, accommo-
dating various reasons for differentially restricting access to fields of
information within them. 313 With the stakes raised by radical changes
in access and dissemination, we have argued in favor of imposing ap-
propriate constraints to compensate for these changes. Although ad-
ditional constraints constitute a change in practice, the goal of such
change is to sustain the underlying interests and values at stake. We
may be defying convention but we do so for good reason, just as a
good general dares not apply conventional strategy designed for a
world with cannons to a world with ballistic missiles.
But the departure from convention we have suggested, particu-
larly in a system that has evolved over centuries and whose very labe-
ling of cases inextricably binds them to parties’ names, calls for care-
ful attention to counter arguments. We have considered the interests
of parties and other participants and we have considered goals and
values of the context; neither of these offers overwhelming counter-
weight. Although convention itself surely counts for something, resis-
tance to change that would comprehensively block or diminish access
to named individuals is rooted in something else, another factor that
has not been fully acknowledged nor reckoned with.
Among the staunchest supporters of unrestricted online access
are data aggregators, such as Choicepoint and LexisNexis. Although
they align themselves with supporters of openness, transparency, citi-
zen oversight, and accountability, data aggregators mine personal in-
formation in public, including court records, for different reasons en-

311. See THE SEDONA GUIDELINES: BEST PRACTICES ADDRESSING PROTECTIVE ORDERS,
CONFIDENTIALITY & PUBLIC ACCESS IN CIVIL CASES 46–49 (Ronald J. Hedges & Kenneth J.
Withers eds., 2005) (providing an overview of “Privacy and Public Access to the Courts in
an Electronic World”).
312. See, e.g., Mulvaney, supra note 93 (reporting ACLU opposition to a proposed
change to a rule “that would allow people to seek to block certain information from ap-
pearing in court records available over the Internet”).
313. See supra Part V.A (discussing “green,” “red,” and “orange” records).
846 MARYLAND LAW REVIEW [Vol. 71:772

tirely. 314 Their business is assembling into dossiers as much informa-


tion as possible about individuals, touting the utility of this service to
all of hardworking, law-abiding society. 315 Is it not useful to know
whether a neighbor has committed fraud, a job applicant has embez-
zled funds, a suitor has undergone a nasty divorce, a political candi-
date has been sued for business corruption, or an insurance applicant
has been convicted of drunk driving? But it is important to notice
that something new has edged into the landscape. These reasons in-
voke a different value in support of faster, cheaper, and uncon-
strained access to full court records online having little to do with
openness, transparency, citizen oversight, and accountability. By their
light, court records are valuable because they constitute a repository
of useful information about people.
As useful as court records may be as sources for aggregated repo-
sitories of information about people, we are not convinced that an ar-
gument has yet been made that prioritizes this function in shaping
access policies to court records. Indeed, explicit acknowledgment of
it, evaluation of its legitimacy as a policy driver, and its relative benefit
weighed against other goals and values, has been conspicuous in its
absence. In bringing to light this “hidden variable,” we recognize as
precursor Justice Holmes’s justification of open records,
not because the controversies of one citizen with another are
of public concern, but because it is of the highest moment
that those who administer justice should always act under
the sense of public responsibility, and that every citizen
should be able to satisfy himself with his own eyes as to the
mode in which a public duty is performed. 316
In other words, satisfying with our “own eyes” that public duty is being
properly performed is primary; prying into the private disputes of
other citizens is a mere artifact of a system that must deliver the
second with the first. The radical option might have satisfied Justice
Holmes well.
Admittedly, we have not settled the question of whether court
records should function as a repository of personal information for use
by government or corporate actors as a basis for getting to know or

314. See, e.g., Robert O’Harrow Jr., In Age of Security, Firm Mines Wealth of Personal Data,
WASH. POST, Jan. 20, 2005 at A1 (reporting on Choicepoint’s private intelligence activi-
ties).
315. See DANIEL J. SOLOVE, THE DIGITAL PERSON: TECHNOLOGY AND PRIVACY IN THE
INFORMATION AGE 16–26 (2004) (providing an overview of the history of private-sector da-
tabases).
316. Cowley v. Pulsifer, 137 Mass. 392, 394 (1884).
2012] ONLINE COURT RECORDS 847

vetting people of interest. Instead, we have endeavored only to ex-


pose the role that this function plays in influencing the shape of poli-
cy. The question deserves close and explicit scrutiny as a public mat-
ter. Should it be judged legitimate for court records to function as
repositories, the next urgent matter would be to articulate the courts’
responsibilities as creators of databases of personal information: Is the
role curatorial? Are courts trustees and guardians? What is the extent
of their obligation to assure accuracy and security? If information in
records is decisively affecting people’s lives during or after the com-
pletion of court cases, these matters are crucial to settle. In defining
the role of courts in relation to third-party information intermedia-
ries, it should also be discussed, as a public matter, what the respective
obligations are of both parties—courts and commercial consumers of
court records—in maintaining, sharing, and using information in
records.